2. 1. hi .Uu:
$4.“... "1...“; .
. , . . . . 4 . 4 :35- .
x .x... . .1... . . , .
. » 5.

:vvv...fl.i .3:

Iii... 1..

u..- 36.3%....0
£712..

. a»:
A .t’

.. ..

.39....“ .5

. s.
.2.
... .x .i 3
.73. L\\ud5.A
J.

511;): to.

rm.‘

1 ‘11;
K

..
u i. ‘5
u....§..k.n
1.. 5....

..?.v..
.1. . 3......
4.4.x 508......

. . & iapuf. .
1.... :35.

.33.:
5‘11
1..
3.. .. z I. . 21.... . an...
s}; . 3 . .~.n.....n....a . v
3...!
3:41;
2. .

t... 32...:
«.2 ‘11:...‘1.

.xrv).yv
.05.

t f.) . ....
C. i!
.2 . ‘
15... $3....
u;

1|. «lyv I .
r 1V.xl. );
It... .2: .

K. it
35:12::

       

    

 

r)..." .

.iFV. ;.
. 1. it;

 

. . .1.» 5..
w 5...; .

  

m... Ilfllilliillllifliifitifllflfli'lllliiifllfiil

3 1293 00892 7265

This is to certify that the
dissertation entitled

Differential selection and adaptation in different
host environments: genotypic and phenotypic variation
in host use traits in the tiger swallowtail butterfly,
Papilio glaucus L.
presented by

 

Janice L. Bossart

has been accepted towards fulfillment
of the requirements for

PhD degreein Entomology and _
Program in Ecology and Evolutionary Biology

/ WM

Major professor

Date 8 June 1993

 

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

LIBRARY
Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or betore date due.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution
cmmmuj

DIFFERENTIAL SELECTION AND ADAPTATION IN DIFFERENT
HOST ENVIRONMENTS: GENOTYPIC AND PHENOTYPIC VARIATION
IN HOST USE TRAITS IN THE TIGER SWALLOWTAIL BUTTERFLY,

PAPILIO GLAUCUS L.

By

Janice L. Bossart

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCI‘OR OF PHILOSOPHY

Department of Entomology
and
Program in Ecology and Evolutionary Biology

1993

ABSTRACT _
DIFFERENTIAL SELECIION AND ADAPTATION IN DIFFERENT
HOST ENVIRONMENTS: GENOTYPIC AND PHENOTYPIC VARIATION
IN HOST USE TRAITS IN THE TIGER SWALLOWTAIL BUTTERFLY,
PAPILIO GLA UCUS L.

By

Janice L. Bossart

Host plant use by insect generalists is a function of the insect’s ability to
behaviorally recognize and physiologically use different host species. Hence, the
evolution of host associations in phytophagous insects will depend upon the
extent of adaptively significant genetic variation for these behavioral and
physiological traits, and the factors maintaining this variation. Evaluating the
extent of genetic variation for host preference and larval tolerance in the
polyphagous, tree-feeding species, Papilio glaucus, the eastern tiger swallowtail
butterfly, was the impetus for the research presented in this dissertation.
Specifically, I examined whether Ohio, Georgia and Florida P. glaucus
populations and families within populations responded differently to different
host species, and whether the response was genetically based and adaptively
significant. I compared patterns observed for this ecologically important
variation with those observed for electrophoretically detectable variation.

Adaptively significant genetic variation for host preference and larval
tolerance was present among and within P. glaucus populations. Geographic
populations expressed differing abilities to recognize and use locally abundant
hosts, presumably as a result of selective pressures imposed by these hosts.

Significant interactions were present between host species and families,

establishing that polyphagy in P. glaucus is a function of a collection of
differentially adapted genotypes, rather than a single, particularly robust
genotype. Physiological tradeoffs in the ability to use M. virginiana and
L. tulipifera may, in part, underlie this differential performance.

The geographic differentiation observed for larval tolerance and host
preference contrasted with the lack of any such pattern for electrophoretically
detectable, presumably neutral variation. The electrophoretic results indicated
that gene flow is common among P. glaucus families and populations. The lack
of optimization across hosts, despite ample opportunity, suggested that
differential selection on ecologically important variation in P. glaucus will
continue to be countered by gene flow, thereby maintaining genetic variation for
host-use traits. In the absence of more substantial barriers to gene flow, the
eastern tiger swallowtail will continue to be comprised of a mosaic of genotypes

exhibiting differing abilities to use different host species.

ACKNOWLEDGEMENTS

I extend my sincere thanks to Jim Maudsley and family for their
assistance in the collection of butterflies from Georgia, and to Mary Jeanne and
Lloyd Johnson for their willingness to collect and mail host foliage when
needed. Thanks also to my mom and dad for providing a base station, complete
with home cooked meals, during my research forays to southern Ohio. Dad, you
became quite the able bodied research associate. The electrophoretic expertise
of Doozie Snider and her willing assistance in the laboratory are greatfully
acknowledged. This project was tremendously enriched by the many hours of
stimulating, intellectual discussion with Matt Ayres, Guy Bush, Bob Hagen, Don
Hall, Kelly Johnson, Bob Lederhouse, James Nitao, Mark Scriber, Mike Singer,
Fred Stehr and Steve Tonsor. Financial support was provided by NSF BSR
9001391, national and local chapters of Sigma Xi, and graduate fellowships from
the College of Natural Science and Department of Entomology at Michigan
State University. I single out my daughter, Allyson, for special recognition. It

was tough at times, Allyson, but we did it.

iv

TABLE OF CONTENTS

LIST OF TABLES ......................................... viii
LIST OF FIGURES .......................................... x
GENERAL INTRODUCTION .................................. 1

MANUSCRIPT 1 - Maintenance of ecologically significant genetic
variation in the tiger swallowtail butterfly

through differential selection and gene flow .......... 5

Abstract ............................................. 5
Introduction .......................................... 6
Materials and Methods .................................. 9
General ......................................... 9
Oviposition preference ............................. 11
larval nutritional physiology ......................... 12
Electrophoretic analyses ............................ 13
Results ............................................. 15
Oviposition preference ............................. 15
Larval nutritional physiology ......................... 15
Electrophoretic analyses ............................ 22

Discussion ........................................... 24

Interpretation of variation in ecological traits ............. 24
Interpretation of electrophoretic variation ............... 26
Differential selection vs. gene flow .................... 27

MANUSCRIPT 2 - Genetic and nongenetic components of oviposition
preference variation within tiger swallowtail

butterfly populations .......................... 31

Abstract ............................................ 31
Introduction .......................................... 32
Materials and Methods .................................. 35
General ........................................ 35
Assessment of phenotypic variation .................... 37
Genetic component of preference variation ............... 38
Non-genetic component of preference variation ............ 39
Statistical anaylses ................................. 41
Results ............................................. 42
Phenotypic variation ............................... 42
Genetic component of variation ....................... 44
Non-genetic component of variation .................... 46
Discussion ........................................... 52

MANUSCRIPT 3 - Intrapopulation genetic variation and covariation
in larval performance across hosts in the

polyphagous eastern tiger swallowtail ., ............. 64
Abstract ............................................ 64
Introduction .......................................... 64

Materials and Methods .................................. 68

Larval sources .................................... 68

Larval performance ................................ 69

Statistical analyses ................................. 70

Results ............................................. 71
Discussion ........................................... 85
GENERAL CONCLUSIONS .................................. 90
APPENDIX - Record of deposition of voucher specimens .............. 93
BIBLIOGRAPHY ........................................... 95

vii

LIST OF TABLES

Table 1. Mixed model analysis of variance of 1988 larval
performance data comparing P. glaucus larvae from Ohio,
Georgia and Florida populations on L. tulipzfera,
M. virginiana and P. serotina foliage; ns = not significant at
= .05 .......................................... 17

Table 2. Mixed model analysis of variance of 1989 larval
performance data comparing P. glaucus larvae from Ohio,
Georgia and Florida populations on L. tulipifera,
M. virginiana and P. serotina foliage; ns = not
significant at p = .05. The fami1y(pop) component
could not be estimated because not all families were
represented across all hosts ........................... 20

Table 3. Wright’s F-statistics for P. glaucus population samples
from Ohio, Georgia and Florida. N=48-74, except for
the X-linked loci, P3GDH and TPI, where N=23-25.
Means and standard errors estimated by jackknife
procedure over loci (Sokal and Rohlf 1981). Negative
values of PST interpreted as F5).r = 0 ..................... 23

Table 4. Analysis of variance comparing larval duration for full-
sibling P. glaucus families collected from different
populations and broods and reared on P. serotina,
M. vitginiana or L. tulipzfera ........................... 72

Table 5. Analysis of variance comparing relative growth rate for
full-sibling P. glaucus families collected from different
populations and broods and reared on P. serotina,
M. Virginiana or L. tulipzfera ........................... 73

Table 6. Analysis of variance comparing pupal mass for full-sibling
P. glaucus families collected from different populations
and broods and reared on P. serotina, M. vuginiana or
L. tulipifera ....................................... 74

Table 7.

Table 8.

Correlation matrix for P. glaucus populations and broods.
Product-moment correlations are of larval host use traits

expressed by full-siblings reared in different host

environments. PSday, MVday and LTday = larval

duration on P. serotina, M. virginiana and L. tulzpifera,

respectively; PSwgt, Mngt and Lngt = pupal mass on

P. serotina, M. virginiana and L. tulipifera, respectively;

PSrgr, MVrgr and LTrgr = relative growth rate on

P. serotina, M. virginiana and L. adiptfera, respectively ........ 83

Correlation matrix of full-sibling larval survival across 3

host species for P. glaucus populations and broods. PSsrv,

MVsrv and LTsrv = survival on P. serotina, M. vuginiana

and L. tulipifera, respectively .......................... 84

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

LIST OF FIGURES

Frequency distributions of L. tulipifera preference for
Ohio, Georgia and Florida P. glaucus females. Preferences
were assayed for L. tulipifera and M. virginiana in 2-choice
host trials. Bars to the right on the x-axis indicate
increasing L. tulipifera preference, bars to the left, indicate
increasing M. virginiana preference. N = 39 Ohio, 40

Georgia and 32 Florida females tested .................

Mean (1 SE) relative growth rates (RGR) for Ohio,
Georgia and Florida P. glaucus larvae reared on

L. tulipifera, M. virginiana and P. serotina. Ten

to twelve full-siblings from 13 to 23 families

(5 to 16 each year) were allocated to each of the hosts .....

Correlation between pupal mass and larval duration for
Ohio and Florida P. glaucus larvae reared on L. tulipifera
and M. vilginiana. Each point represents the mean of 3 to
30 full-siblings. The analysis was performed on data

pooled from 1988 and 1989 .........................

Distribution of preferences expressed by P. glaucus females
collected from north central Georgia. Distributions reflect
relative preference for L. tulipifera. Preferences were
assessed in 2-choice trials comparing L. tulipifera and one
of four alternate hosts: P. serotina, M. virginiana,

.. 16

.. 19

F. americana or P. mfoliata ............................ 43

Regression of L. tulipifera preference of mothers on mean
L. tulipifera preference of daughters. Solid lines are
regressed through data from different years; the dotted

line is regressed through data from both years combined . . . .

Correlation between L. tulipifera preference of mothers
and the mean relative growth rate of her progeny on

.. 45

L. tulipifera and M. virginiana .......................... 47

Figure 7.

Figure 8.

Figure 9.

Figure 10.

Figure 11.

Figure 12.

Figure 13.

Figure 14.

Figure 15.

Specificity for preferred hosts when compared against a

less preferred host as a function of the wear condition

class of females. Wear condition is a reliable indicator of

relative age. See text for definition of preference index ........ 48

Relative degree of preference exhibited by females as a

function of female fecundity and pupal mass. Females

were tested in L. tulipzfera/M. vizginiana 2-choice trials.

Solid lines are regressed through data from different hosts;

the dotted line is regressed through data from both hosts

combined. See text for definition of preference index ......... 50

Relative degree of preference exhibited by females as a

function of egg load. Females were tested in

L. tulipifera/M. virginiana 2-choice trials. Solid lines

are regressed through data from different hosts; the dotted

line is regressed through data from both hosts combined.

See text for definition of preference index .................. 51

Distribution of preferences for L. tulipifera expressed by a
population of P. glaucus females as a function of larval
host substrate ...................................... 53

Mean (2 SE) preferences for L. tulipifera and P. serotina
expressed by full-siblings as a function of larval host
substrate ......................................... 54

Reaction norms depicting two types of patterns that can

underlie significant genotype x environment interactions: a)

the rank order of phenotypes on each host is the same,

tradeoffs are absent, b) the rank order of phenotypes is

different on each host, tradeoffs may be present ............ 67

Reaction norms of relative grth rate for full-sibling,
Florida P. glaucus larvae developing on P. serotina,
M. virginiana or L. tulipifera ........................... 76

Reaction norms of relative growth rate for full-sibling,

Georgia P. glaucus larvae developing on P. serotina,

M. vizginiana or L. tulipifera. The use of the symbol

indicates a family not represented on L. tulipifera ........... 77

Reaction norms of relative grth rate for full-sibling,

Ohio P. glaucus larvae developing on P. serotina,

M. virginiana or L. tulipifera. The use of the symbols

indicates families not represented on M. virginiana ........... 78

Figure 16. Reaction norms of percent survival of full-sibling, Florida
P. glaucus larvae developing on P. serotina, M. virginiana or
L. tultpifera ....................................... 80

Figure 17. Reaction norms of percent survival of full-sibling, Georgia
P. glaucus larvae developing on P. serotina, M. virginiana or
L. tulipifera ....................................... 81

Figure 18. Reaction norms of percent survival of full-sibling, Ohio

P. glaucus larvae developing on P. serotina, M. virginiana
or L. tulipifera ..................................... 82

xii

GENERAL INTRODUCTION

Host plant use by insect generalists is a function of both physiological
and behavioral traits, which are, in turn, governed by both genetic and nongenetic
components. Ecological factors exerting selective pressure operate only on
genetic variance. Yet, little is known regarding the extent of interspecific genetic
variation for host preference and larval performance, and even less, regarding
intraspecific genetic variation (Futuyma 1991, Futuyma and Peterson 1985).

Genetic variation for host-use traits is generally extensive in generalist
herbivore populations. This is not surprising given the multitude of selective
factors associated with different host species that shape patterns of host-use in
insects. Different hosts present unique selective environments, reflecting
differences in physical and chemical composition (Fox and Morrow 1981),
associated natural enemies or competitors (Bernays and Graham 1988, Strong et
a1. 1984), microclimates (Grossmueller and Iederhouse 1985), and abundance
and distribution (Gilbert and Singer 1975). The probability that a single genotype
is able to respond optimally in all situations seems remote. That different
genotypes are optimal in different situations seems much more likely. Such
differential selection in an ecologically diverse, stochastically changing
environment, is thought to explain the persistence of much of the genetic

variation for host-use traits observed in natural populations (Hedrick et a1. 1976,

2
Hedrick 1986, Koehn and Hilbish 1987, Powell and Taylor 1979). Genotypes
selectively eliminated on one host species or in one population will persist as long
as gene flow occurs among populations and mating is mostly random within
populations. That differential selection, in part, serves to maintain genetic
variation seems even more likely given that most host-use traits are probably
polygenic.

The amount of genetic variation harbored by a species determines the
ability of the species to adapt to environmental conditions that vary through space
and time. Evaluating the extent of genetic variation in traits affecting the use of
different hosts and the factors maintaining this variation is a necessary requisite
of studies examining the evolution of host associations in phytophagous insects.
Such information is especially important with regard to species that use multiple
hosts or span a large geographic area, since environmental heterogeneity is
greatest for these species.

The tree-feeding Papilio spp. group is perhaps the most intensively
studied group of generalist herbivores to date. All members of this group,

P. glaucus, P. canadensis, P. eutymedon, P. alexiares, P. rutulus and

P. multicaudatus, are characterized by their ability to feed on several host families
(Scriber et al. 1991). Genetically based variation in host associations appears to
be a major factor delimiting the geographic range of this group (Scriber and
Lederhouse 1992). The abundance of information gathered on the Papilio spp.
group far exceeds that for most other generalist insects. Yet even in this well
studied group, the extent of genetic variation for host-use traits among

populations within a species (although see Scriber 1986), and among individuals

3
within populations remains virtually unknown. Most initial studies focused on
interspecific variation and, in particular, on interspecific variation between
P. canadensis and P. glaucus, two closely related sister taxa. The discovery of
genetically based differences in host use between these two species (Lindroth et
al. 1986, Scriber 1982, 1988, Scriber et al. 1989), in combination with differences
in diapause physiology (Rockey et al. 1987a, 1987b), color polymorphisms (Scriber
1988, Hagen and Scriber 1989), wing morphometrics (Luebeke et al. 1988),
allozymes (Hagen and Scriber 1991), and mitochondrial restriction sites (Sperling
1991), recently led to the designation of P. canadensis and P. glaucus as distinct
species (Hagen et al. 1991). These sister taxa are prime examples of species that
are subjected to extensive environmental heterogeneity. In combination,
P. glaucus and P. canadensis utilize hosts from approximately 9 plant families and
their geographic range covers much of North America. The transition zone
between the two species occurs through central Wisconsin and Michigan and
extends eastward through central New York and southern New England (Scriber
1988), corresponding with the transition between boreal coniferous forests to the
north and mixed deciduous forests to the south.

P. glaucus, the eastern tiger swallowtail butterfly, is especially
amenable to an analysis of the interplay among environmental heterogeneity,
differential selection, and gene flow and random mating within a single species.
P. glaucus is the most polyphagous member of the Papilio spp. group, using hosts
from at least 7 different plant families. No single host species is coincident with
the geographic range of P. glaucus, an area which extends south from Michigan to

Florida, west to Texas, then south into Mexico. Although most P. glaucus

4

populations encounter multiple hosts species, others are virtually restricted to a
single local host.

The issues outlined in the preceding paragraphs served as the impetus
for the following dissertation. Specific objectives included establishing whether
P. glaucus populations and families within populations responded differentially to
different host species, whether this response was genetically based and adaptively
significant, and whether gene flow was common. These objectives were addressed
with regard to two important host use traits, larval tolerance and female
preference, for P. glaucus populations from southern Ohio, north central Georgia,
and southern Florida. Specific populations were selected based on the probability
that they represented "pure" P. glaucus, their location along a north-south
transect, and the local abundance of host species. The dissertation is presented
as three separate manuscripts. The first manuscript addresses the hypothesis of
differential selection and gene flow, the second, is a comprehensive analysis of
the genetic and nongenetic components of female host selection, and the final
manuscript examines intrapopulation genetic variation and covariation in larval

performance across hosts.

MANUSCRIPT 1
Maintenance of ecologically significant genetic

variation in the tiger swallowtail butterfly
through difierential selection and gene flow

Abstract

Differential selection in a heterogeneous environment is thought to
promote the maintenance of ecologically significant genetic variation. Variation
is maintained when selection is counterbalanced by the homogenizing effects of
gene flow and random mating. In this study, we examine the importance of
differential selection and gene flow in maintaining genetic variation in Papilio
glaucus. Differential selection on traits contributing to successful use of host
plants (oviposition preference and larval nutritional physiology) was assessed by
comparing the responses of southern Ohio, north central Georgia and southern
Florida populations of P. glaucus to three hosts: Lin'odendron tulipifera, Magnolia
vitginiana and Pmnus serotina. Gene flow among populations was estimated
using allozyme frequencies from 8 polymorphic loci.

Significant genetic differentiation was observed among populations
for both oviposition preference and larval performance. This differentiation can
be interpreted as the result of selection for enhanced use of Magnolia, the
prevalent host in Florida, by Florida P. glaucus. In contrast, no evidence of

population differentiation was revealed by allozyme frequencies. Fgr values

6
were very small and Nm, an estimate of the relative strengths of gene flow and
genetic drift, was large. Results suggest that most of the electrophoretically
detectable variation is neutral and that gene flow among P. glaucus populations
is common. The contrasting patterns of spatial differentiation for ecologically
important variation and lack of differentiation for neutral variation implies that
differential selection among populations will be conterbalanced by gene flow,

thereby maintaining genetic variation for host use traits.

Introduction

Natural insect populations are commonly comprised of an abundance
of different genotypes. In many cases, these different genotypes reflect heritable
differences for ecologically significant traits (e.g., Gould 1979, Jaenike 1989, Via
1984). It is unclear how such non-neutral variation persists in natural
populations. Much of the variation (i.e., that which is nonadaptive) should be
selectively eliminated as a population becomes increasingly adapted to local
conditions. One hypothesis suggests that differential selection in a
heterogeneous environment maintains selectively important genetic variation
(Gillespie 1973, Gillespie and Turelli 1989, Hedrick et al. 1976, Levene 1953,
Nevo 1988, Powell and Taylor 1979, Via and Lande 1985). In part, this
hypothesis is based on the assumption that different genotypes are more or less
fit under different environmental conditions and that no single genotype is
optimal under all conditions (Powell 1971, Powell and Taylor 1979). Such
differential selection can preserve genetic variation in a population when it is

offset bythe homogenizing effects of gene flow and random mating (Futuyma

7
1986). A genotype selectively eliminated from a population under unfavorable
conditions can be reintroduced to the gene pool as a result of gene flow from
favorable environments and random mating among individuals (Koehn and
Hilbish 1987). As a consequence, a balance may be achieved between selection,
which acts to reduce genetic variation, and gene flow and random mating, which
serve to counteract this loss (Slatkin 1973).

A major determinant of environmental heterogeneity for
polyphagous herbivores encountering an array of plant species is host diversity.
The physical and chemical attributes of different plant species can be highly
variable (Rosenthal and J anzen 1979, Juniper and Southwood 1986) and
herbivores with a particular genotype are frequently affected differently by
individual host species. As such, different plants can be regarded as separate
environments. The extensive variability in host use traits for herbivores feeding
on an array of plant species suggests that host diversity is indeed an important
determinant of genetic variation (e.g., Futuyma and Peterson 1985, Nitao et al.
1991a, Rausher 1984, Rossiter 1987, Scriber 1986, Via 1984). However, the
extent to which this variability might be the consequence of differential selection
on different hosts and gene flow among these different environments remains
largely uninvestigated.

Papilio glaucus L., the eastern tiger swallowtail butterfly, is well
suited for examining the importance of host diversity and gene flow in
maintaining genetic variation in host use traits. It is exceptional in its breadth
of diet, using at least 18 different host species from 7 plant families (Bossart and

Scnber 1993, Scriber 1984). However, separate populations and even

8
individuals within a population encounter only a subset of the potential hosts
available (Scriber 1983, 1986). No single host has a range which completely
overlaps that of the butterfly. Moreover, the relative abundance of each host
varies from site to site within its distribution. Previous studies documented
variation in response to hosts by P. glaucus and P. canadensis (Scriber 1988,
Scriber et al. 1989), and in part, lead to these former conspecifics being
regarded as distinct species (Hagen et al. 1992). More recent studies
documented variation in host use among populations within P. glaucus (Scriber
1986). However, these P. glaucus populations were not tested simultaneously
and it is unclear whether observed variation is genetically based or a reflection
of environmental effects, such as seasonal variation in host quality.

In the current study, we investigated whether P. glaucus populations
experience differential selection on locally abundant host species and, if so,
whether this differential selection has resulted in genetic divergence among the
populations. We contrasted the patterns observed for variation in ecologically
important traits with those obtained from electrophoretic analyses of loci
ostensibly reflecting selectively neutral variation (Karl and Avise 1992). Our
specific objectives were to 1) quantify genetic variation in P. glaucus for two
ecologically important host use traits, oviposition preference and larval
nutritional physiology; 2) determine whether genetic differentiation for these
traits has occurred among three geographically distant P. glaucus populations
located along a north-south transect; and 3) use electrophoretic data to estimate

rates of gene flow and examine population substructuring.

9

A secondary interest was to determine if differentiation among
populations was the result of local adaptation. Although demonstrating
adaptation directly is difficult, it’s possible to make certain predictions regarding
the behavior of host use traits in an adapted versus nonadapted population. We
predicted that if locally abundant hosts are selective agents effecting
differentiation among P. glaucus populations, then development time and pupal
mass would be positively correlated on local hosts but negatively correlated or
uncorrelated on rare hosts. Such a result would be expected if adapted
populations are mostly comprised of uniformly well adapted genotypes, while
nonadapted populations are comprised of a mixture of preadapted and

nonadapted genotypes.

Materials and Methods

General

Butterfly Sources: Adult P. glaucus butterflies were collected from
field populations in southern Ohio (Lawrence Co.), north central Georgia
(Clarke Co.) and southern Florida (Highland Co.). These populations span
1300 km, each separated from the nearest sampled population by approximately
650 km. Butterflies were collected from multiple sites in each region. Sites
were separated by no more than 30 km and no less than 4 km. Field collected
butterflies were placed in individual glassine envelopes, then transported on ice
or shipped using overnight delivery to our laboratory at Michigan State
University. All field collected butterflies not used for oviposition were frozen at

-80°C to preserve tissues for allozyme electrophoresis. Females to be used to

10
generate eggs were fed a 20% honey solution immediately upon arrival, then
kept at 24°C for 24 h. This procedure provided for maximum survival of
ovipositing females. Females were individually placed in (10 cm x 20 cm x 27
cm) clear plastic ”shoeboxes" (Tristate Plastics) with sprigs of appropriate foliage
and fed a 20% honey solution daily. Boxes were maintained under artificial
illumination, alternating 4h:4h photozscotophase. This photoperiod permitted
maximum oviposition, while preventing high mortality due to overheating. Eggs
were collected daily and resulting neonate larvae used for the nutritional
physiology studies and to generate adults for the oviposition assays.

Foliage: Responses to Magnolia virginiana L. and Lin'odendron
tulipifera L. (Magnoliaceae), and Prunus serotina Ehrhart (Rosaceae), foliage
were compared. All three hosts support generally high levels of survival and
larval growth performance. The frequency of use of these three hosts is
different among the three P. glaucus populations. Florida P. glaucus are largely
restricted to M. virginiana, the only common host throughout much of peninsular
Florida (Scriber 1986). In contrast, Georgia and Ohio populations rarely or
never encounter M. virginiana but frequently encounter L. tulipifera and
P. serotina.

L. tulipifera and P. serotina foliage was collected at least every third
day from various areas in the vicinity of the Michigan State University campus
and stored at 7°C. M. virgimana foliage was collected daily from potted trees
maintained on campus. Foliage sprigs presented to developing larvae and
ovipositing females were placed in water-filled, rubber-capped, plastic vials to

maintain leaf freshness.

11

Oviposition preference

Oviposition preferences were assayed using progeny from field
collected females. To minimize possible maternal effects, larvae were fed a
common host (P. serotina). After emergence, females were hand paired with
males originating from the same population. Mated females were fed, then kept
at 24°C for 24 h prior to testing. Emergence within populations spanned a three
week period from mid-July to mid-August, dictating the period over which the
oviposition trials were conducted.

Preferences were assessed for two of the three host species,
M. vuginiana and L. tulipifera. Assays comparing P. serotina with the other hosts
were not included because preliminary studies indicated that females from all
populations rarely oviposited on P. serotina in choice tests when alternate hosts
were available. Fresh host sprigs of similar leaf surface area were positioned in
opposite corners along the long side of the box towards the lights. Females
were free to move within the box and were commonly observed fluttering
between host sprigs. The position of each sprig was alternated once a day to
control for positional effects and any wilted foliage was replaced. The position
of individual boxes in relation to the light source was randomized at each
feeding. After 4 days, the total number of eggs deposited on each host species
was counted. The few stray eggs that were placed on the paper lining or box
were generally adjacent to a particular leaf but nonetheless excluded from our
analyses. Preferences were calculated as the percentage of total eggs oviposited
on each host over the 4-day trial. (Compared to an analysis of daily

percentages, this technique is less sensitive to low daily egg numbers). Only

12
females ovipositing at least 25 eggs were included in the analysis; 39 Ohio

females, 40 Georgia females and 32 Florida females met this criteria.

Larval nutritional physiology

The performance of larvae from each of the three populations was
compared on each of the three hosts in two separate studies. In 1988, the
performance of offspring of field-collected females was assessed. Due to
differences in emergence times among populations, Florida larve could not be
tested concurrently with Ohio and Georgia larvae. In 1989, we analyzed the
performance of progeny generated from laboratory-reared females fed a
common host (P. serotina) during their development.

A similar protocol for testing larvae was followed in both studies.
Eggs were checked for eclosion every 2-3 h during the day. Resulting neonate
larvae were weighed to the nearest mg, then randomly selected and distributed
across the three host species. Ten to twelve full-siblings from 13 to 23 families
from each population (5 to 16 each year) were allocated to each of the hosts.
Larvae were reared individually to pupation in (150 cm x 25 cm) screened,
plastic petri dishes containing the appropriate host foliage (24°C, 18h:6h
photozscotophase). Larvae were checked daily. Fresh foliage was provided at
least every other day. Larval duration, pupal mass and sex were recorded for
each individual. Larval duration was defined as the period from day of eclosion
to the prepupal stage, whereupon larvae cease feeding, void gut contents and
undergo a conspicuous color change from green to brown. Pupae were collected

and weighed 24 h after shedding their larval exoskeleton, then placed in

13
individual screen cages until adult emergence. Relative growth rate (RGR) was

calculated as,

RGR = (WP - Wn/«WP- W2) x D). where
(W,) is initial larval mass,
(W,) is pupal mass, and

(D) is larval duration.

Electrophoretic analyses

Of the 26 consistently resovable allozyme loci in P. glaucus (Hagen and
Scriber 1991), twelve are polymorphic and useful for analyzing population
substructuring. Loci were examined using electrophoresis on thin-layer cellulose
acetate plates (see Table 3). Approximately 1/4 to 1/3 of butterfly abdomen
was homogenized in 300 pl of extraction buffer and then centrifuged for 8
minutes at 16,000 x g. The resulting supernatants were electrophoresed on
cellulose acetate plates (Helena laboratories; Beaumont, Texas). Buffer
composition, electrophoretic conditions and staining procedures were adapted
from Richardson et al. (1986) and Harris and Hopkinson (1978) and follow
those of Hagen and Scriber (1991). Alleles were designated according to their
relative mobility; negative numbers were assigned to cathodally migrating
allozymes. Three standards (i.e., butterflies previously scored for genotype on
earlier runs) were included on subsequent plates to aid in scoring. As an
additional check, every fifth plate was a rerun of 12 randomly chosen individuals

from the 4 previous plates (3 from each).

14

Wright’s F-statistics were estimated for each locus (Wright 1951, Weir
and Cockerham 1984, Long 1986). Fgr measures variance in allele frequencies
among demes relative to the expectation when genes are randomly assorted
among populations. Frr measures deviations from homozygote frequencies
expected in a panmictic population by estimating inbreeding within individuals
relative to Hardy-Weinberg expectations for the total population. F rs measures
homozygote frequency relative to Hardy-Weinberg expectations within
populations; positive values indicating a deficit of heterozygotes, negative values
indicating an excess of heterozygotes. For neutral alleles, Fgr provides an
indirect mechanism for estimating the degree of subdivision between populations

and thus the relative strengths of gene flow and random drift (Nm):

(Nm)est = (l/FSI‘ ' 1)/4

where N is the effective size of a population and m is the number of migrants
(Slatkin 1985, 1987).

All indirect and direct methods currently available for assessing gene
flow have associated assumptions and drawbacks and often generate disparate
estimates (Slatkin 1985, 1987; Johnson et al. 1988, Whitlock 1992). To
compound the problem, patterns of population substructuring revealed from
electrophoretic analyses, the classical and most widely used technique, may
contradict patterns revealed from other molecular techniques (Karl and Avise
1992). Despite these problems, Fgr values can provide a conventional means of
estimating gene flow for comparison with published values. It remains the most

widely used method for estimating gene flow.

15
Results

Oviposition preference

Genetic differentiation in preference for L. tulipifera and M. virginiana
was evident among the P. glaucus populations (Figure 1). Ohio and Georgia
females were not only less willing than Florida females to oviposit on M.
virginiana, the prevalent host in Florida, but clearly preferred L. tulipifera, a host
commonly encountered in Ohio and Georgia. Approximately half of the Ohio
and Georgia females tested (44 and 48%, respectively) oviposited greater than
70% of their eggs on L. tulipifera, evidenced by distributions skewed to the right.
In contrast, few of the Florida females tested (9%) oviposited greater than 70%
of their eggs on L tulipifera. The majority of Florida females oviposited
relatively equal numbers of eggs on both hosts, evidenced by a distribution
centered around the 51 - 60% category. A Kruskal-Wallis analysis on the ranks
of these preferences confirmed the pressence of a significant population effect
(p=0.0048, df =2), resulting from deviation between the Ohio and Florida
distributions, and the Georgia and Florida distributions (Tukey’s studentized
range test; p=.05, df= 108). The Ohio and Georgia distributions were not

significantly different from each other.

Larval nutritional physiology
1988 Performance Study: Populations were significantly differentiated in
their growth responses. There was an overall population effect on relative

growth rate and on larval duration (Table 1), as well as a significant interaction

16

.3058 8—383 «Eco—m mm 3:3 33.800 9. .320 an u 2 .8380on 333.33%; .2

$5885 8865 £2 2: 9 EB gap—£23 93.33.33 4 €63.85 2865 £53 2: no Emma

05 9 «can 233 “we: 36:0-N E 3333?: .2 can 3%ng 4 .8 30.333 0.83 woos—832m .8383
3.333% .m «ESE can «R800 .320 38 83838.. 3355333 .4 we acoustic hacienda A charm

6.8.3.933» .Qco mmmu N.

 

omA col—m owl—h OhI—m owl—n owl—V cilpn onI—N 0N1: o—V
.3. “omen.“ .nconob .u..no., “own.“ . o a . if... Ammmmm

e o .
o n e
e .

    

 
     

     

  
  

 

 

i
. . . u-voo.-¢v-O--.'.v0Iouioovvuo-Qctouvo-
-.o oonc..acne-uIoooooeoouoeeveoenone-coon-uoeo-
o not...onOue00000.0.0.0000.0.0.0.0..05000000-
.ClIO...OOOOOOOOOOOOOOOOOOQOOOOCOOOOOOIO...
COCO....00....00....OOOCOOOOOOOIOOIOOOOO.
ooo.o.00o00000900000000...0.00.0000.no...
. . .................... . .. .

SO|DLLIGJ jo tueomd

02;: I . mm
Emuooo Z
020

 

 

 

 

on

17

team—2 83 £38.."— x “8: 35 £80..— 8.. 883830.. 88 K 2:.
be... . .3sz as? .83 x coca—sack .05 a8: 08 8888030.. 32 K 2:

giant—R «.03 noun—anon 8.. noun—mace“... 80“ k on m .

 

8.3 an Sum
as. 8.2 :8 8 3.58:5. x .8:
38¢. 8N 33 NN 88:88.8
8. 28 N08 a. so: x 88:58
288. NEN 8.me N 8:
8o. NNN 8.83 N 8888 88:5 353
«no. Sm Sum
m8. NE 8. 8 383:5". x no:
88. Nam 88. NN 38:28.2
8. RN N8. 8. 8: x 8:288
8a. 8.3 98. N .8:
8 N3. N8. N 83288 was 395
88. an Sum
88. N: 88°. 8 38:85 x 8:
H88. 8N 88. NN 38588
8. 5. N88. 8. so: u 8:288
8c. No.8 :8. N 8:
88. 8.8 88. N 88388 as. £320 383
Q 12 m2 m3 ooh—om «8F

 

.8. u m 8 «gang... 8: u a: amaze
33.388. .m 333 333.33g: .2 .E0b833 4 :0 838338 «ESE can «H.800 .030 Bed 085.
3.033% .m @5938 Sac 3§§otoa Ria— mwfl .«o 00:33? me $398 $308 0052 A 030,—.

18

between host species and P. glaucus population for all three traits measured.
For two of these traits, larval duration and relative growth rate, the patterns
were identical; only relative growth rates are depicted since this variable
incorporates both larval duration and pupal mass (Figure 2, 1988). Although
Ohio larvae tended to grow slower than Florida larvae, this difference was most
pronounced on M. virginiana (the prevalent Florida host) and was negligible on
L. tulipifera. Georgia larvae exhibited an intermediate rate of growth on

On L. tulipifera or P. serotina, Ohio larvae tended to grow at the slowest
rate, Georgia larvae at the fastest rate and Florida larvae at an intermediate
rate (Figure 2, 1988). Survival was similar for all three populations on all three
hosts, ranging from 71 to 85% on L. tulipifera, 65 to 80% on M. virginiana and
70 to 85% on P. serotina.

1989 Performance Study: When progeny from lab-reared parents were
assayed in 1989, the patterns of differentiation were similar to those observed in
1988 (Table 2; Figure 2, 1989). Though not as pronounced, a latitudinal cline in
relative growth rate was again observed on M. virginiana, but not on L. tulipifera
or P. serotina. Overall survival was much lower in 1989 but again, was similar
across hosts for all three populations, ranging from 25 to 40% on L. tulzpfera, 21
to 40% on M. virginiana and 17 to 23% on P. serotina.

Correlations between pupal mass and larval duration on local and non-
local hosts were consistent with our earlier prediction regarding expected

patterns in an adapted versus nonadapted population (Figure 3). In Florida

-1
x day )

1

19

 

100 1988 [:1 Ohio

m Geor ia
90 . .8
- Florlda

80‘
70-

 

60-

 

 

 

 

 

 

 

 

 

 

4O

 

100
1989

90-
80-

RGR (mg x g

70-

 
     
     
      
       

    
 

. 1
5’...

vv‘
..9

   

     
  
  
   

O

’9

V
O

60—

 

O
O

 

V

’0

v
.5

O
a

O
O

       

v
Q

929

‘

    
  
  
  
    
    
   

       

9'9

 

  
  
     
     
    
    
     

V

’0

O

50-

O

 

.v
9
.O

.v
O
O.‘

v
O

    
  

O

9'.
.:

 

 

 

 

 

 

 

4O

 

 

 

L. tulipifera M. virginiana. P. serot'i'na.

Figure 2. Mean (1- SE) relative growth rates (RGR) for Ohio,
Georgia and Flordia P. glaucus larvae reared on L. tulipifera,

M. virginiana and P. serotina. Ten to twelve full-siblings from 13 to
23 families (5 to 16 each year) were allocated to each of the hosts.

20

 

 

SAN 82 Sum
88. 8.8 8.88 a 8: N 8838.8
88. 2.9. 2N2: N 8m
88. $8 :82 N 8:288 88:5 3:3
88. § sum
8 8. c8. v 88m x noun—88m
88. 88 E. N .8:
88. 8.8 NNE N 8:28am 82 8.:
N88. 3: Sum
8N8. "NJ. 88. a. 8m N 8:288
8c. NN.8 88. N 8:
28. 8.8 88. N 8:288 08¢ 5320 2.83
a m m: 8 83cm :5

 

.38: :8 888 888882 083 8588 =a

8: 3.282. 38823 on 8: 2:8 808980 $88388 2F .3. u Q 8 883%? 8: u m: “amaze
32.88“. .m 98 3:85»? .2 68.5883 4 N8 macaw—smog «28$ 98 «E800 .oEO Boa

oats 383% N 9:89:00 88 8.88.882“ Eta. owE 8 858., 8 max—«SN .388 8082 .N eBay.

21

1L2

 

M. virginiana

l

2.0
1.8

1.6

1.4

l

1.2

l

l

1.0

 

 

 

0-8 I I I I I l I I I
2.2

2.0 4
1.8 '
1.6

 

 

Pu pal Mass (9)

l
O

l
O
O
1:
II
a:
\l

1.4 -
1.2 - O§0 p=.05
1.0

 

l

0 Florida

0 Ohio
0-8 I I I I I I I I I

22 24 26 28 30 32 34 36 38 40

Larval Duration (Days)

 

 

 

 

Figure 3. Correlation between pupal mass and larval duration for
Ohio and Florida P. glaucus larvae reared on L. ndzpifem and M.
virginiana. Each point represents the mean of 3 to 30 full-siblings.
The analysis was performed on data pooled from 1988 and 1989.

22
P. glaucus, pupal mass and larval duration were positively correlated on
M. virginiana, the local host, but uncorrelated on L. tulipifera. In contrast, in
Ohio P. glaucus these traits were positively correlated on L. tulipifera, the local

host, but negatively correlated on M. virginiana, the host not encountered.

Electrophoretic analyses

Allele frequencies for A02 and MP1 could not be scored consistently
and were excluded from further analyses. Two loci, P3GDH and TPI, are
X-linked, requiring that females (the heterogametic sex in Papilio) be omitted
from F-statistic calculations. Fls and Frr values were small and nonsignificant at
all loci except HBDH and PEP-LA (Table 3). At these two loci, heterozygote
frequencies were lower than expected. All F3.r values were nonsignificant
(Table 3). Negative values of Fgr result from sampling error and were
interpreted as Fsr = 0 (Long 1986). The mean jackknife estimate of F5;r across
loci was 0.00075 when HBDH and PEP-LA were included in the analysis and
0.00091, when excluded, with both estimates indicating a virtual absence of
genetic differentiation among populations. The estimate of FSF excluding
HBDH and PEP-LA is most appropriate for calculating Nm since variation at
these loci may not be neutral (this point is addressed in the discussion). Nm,

using mean Fs'r = .00091 was estimated to be 275.

23

Table 3. Wright’s F-statistics for P. glaucus population
samples from Ohio, Georgia and Florida. N =48-74,
except for the X—linked loci, P3GDH and TPI, where
N=23-25. Means and standard errors estimated by
jackknife procedure over loci (Sokal and Rohlf 1981).
Negative values of Fgr interpreted as Fsr = 0.

 

 

 

LOCUS F15 175,. F”
AAT-l 0.0080 0.0015 0.0095
AC-l 0.0087 0.0033 0.0055
GPI 0.0195 0.0028 0.0224
HBDH 0.1112 0.0052 0.1066
IDH-l 0.0171 0.0056 0.0116
113112 0.0492 0.0069 0.0558
P3GDH 0.0175 0.0119 0.0295
PEP-LA 0.3299 0.0012 0.3291
PGM 0.0046 0.0005 0.0042
TPI 0.0446 0.0103 0.0348
Mean 0.0512 0.00277 0.0466
0.00091*
S.E. 0.0205 0.00169 0.0341
0.00086'

 

‘The calculation of the jackknife estimate of Fgr
(: SE) does not include HBDH or PEP-LA.

24
Discussion
Interpretation of variation in ecological traits

Like many other herbivorous insects, P. glaucus exhibits substantial
variation for ecologically important host-use traits (Ayres et al. 1991, Bossart
and Scriber 1993). At least part of this variation is associated with the
differential use of local host species. Genetic differentiation has occurred
among P. glaucus populations in both oviposition preference and larval
performance, apparently as a result of selection for enhanced recognition of and
performance on a locally abundant host. The similarity between the 1988 and
1989 assays indicates this differentiation is not simply due to maternal or
environmental effects.

Of the three hosts tested, we conclude that selection for enhanced use
of M. virginiana is largely responsible for much of the spatial differentiation.
We draw this conclusion based on three results. First, Florida P. glaucus grew
at a faster rate on M. virginiana (the prevalent Florida host) than either of the
other populations. (It is noteworthy that selection for enhanced use of
M. virginiana did not decrease performance on L. tulipifera, suggesting the
absence of a tradeoff between L. tulipifera use and M. virginiana use). In
contrast neither the Ohio nor Georgia P. glaucus populations exhibited
enhanced performance on L. tulipifera or P. serotina, their locally abundant
hosts. Second, Georgia larvae tended to grow at an intermediate rate on
M. virginiana. This pattern is consistent with what would be expected if a
genotypic cline has formed as the result of gene flow between an adapted

Florida population and a nonadapted Ohio population. Third, Florida

25
P. glaucus exhibited an increased rate of oviposition on M. virginiana. This
pattern is counter to what has normally been observed in Papilio spp. In
general, L. tulipifera is a preferred oviposition substrate, even for those species
in which L. tulipifera foliage is toxic to developing larvae (Scriber et al. 19918.,
1991b). The oviposition patterns exhibited by Ohio and Georgia females seem
to reflect this same general willingness to oviposit on L. tulipifera.

The correlations observed between pupal mass and larval duration
further support our interpretation of M. virginiana as a primary selective agent.
In the Florida population, larval duration and pupal mass were positively
correlated on M. virginiana, reflecting uniform rates of growth among families.
Such was not the case in the Ohio populations where these two traits were
negatively correlated, indicating variation in rates of growth among families and
the presence of genotypes that were not adapted to use this host.

L. tulipifera appears to be a less important selective agent. Only a weak
correlation was detected between larval duration and pupal mass in the
supposedly adapted Ohio population on L. tulipifera. In fact, even a population
not sympatric with tuliptree has little problem using L. tulipifera and females
oviposit willingly on this host. It may be that L. tulipifera is a more acceptable
host in general. A mixture of neolignans, compounds present in M. Virginiana
foliage, are known to decrease performance in unadapted P. glaucus populations,
unadapted Callosamia spp;, the polyphagous Hyphantn'a cunea, Aedes aegypti and
Artemia salina (Nitao et al. 1991b, J. K. Nitao et al. unpubl., K. S. Johnson et al.
unpubl.). These same compounds have not been detected in L. tulipifera foliage

(K. S. Johnson et al. unpubl.).

26
P. serotina also appears to be less important in effecting adaptive
change. Females from all three populations were unwilling to oviposit on
P. serotina in laboratory choice-tests and larvae from all three populations grew
at similar rates on this foliage. These findings need to be interpreted with
caution though since larvae are found on P. serotina in nature. Moreover, other
factors associated with P. serotina besides nutritional quality, e.g. enemy free

space, may be more important as selective agents (Scriber and Lederhouse

1992).

Interpretation of electrophoretic variation

Variation at all loci examined, excluding HBDH and PEP-LA, is
probably neutral and the similarities in frequencies among populations, probably
the result of gene flow. This interpretation is based on the fact that F15, Frr and
Fgr values were small and nonsignificant at all loci except HBDH and PEP-LA,
indicating a lack of departure from Hardy-Weinberg expectations and a lack of
differentiation among populations. To invoke a selective interpretation to
account for this pattern would require that selection operate in such a way as to
mimic Hardy-Weinberg frequencies at these 8 independent, polymorphic loci
and simultaneously generate parallel patterns among all 3 P. glaucus
populations. Such an interpretation seems implausible since selection is much
more likely to produce heterogeneity among loci (Slatkin 1987).

This same interpretation is likely not appropriate for explaining
variation at HBDH and PEP-LA. While Fsr values were again nonsignificant,

F15 values were large and positive, indicating a deficit of heterozygotes within

27
subpopulations. Such departures from expectations could result either from
selection against heterozygotes or non-random mating among genotypes, but
must be indicative of a process that is affecting all 3 populations similarly given
the nonsignificant Fgr values. A tempting interpretation is to conclude that
selection is effecting genotypic frequencies at HBDH and PEP-LA since
non-random mating should impact all 10 loci uniformly. However, non-random
mate choice on some unidentified trait could correlate with genotype at HBDH
and PEP-LA, also resulting in non-uniformity among loci. Determining which
mechanism underlies these patterns is impossible without further study.

Scoring error due to non-detectable variation is a caveat that must also
be considered when interpreting FIs values. Such error is inherent in most
electrophoretic analyzes and may prevent distinquishing heterozygotes from
homozygotes if different alleles migrate to very similar regions. Erroneously
scoring heterozygotes as homozygotes (or vice versa) would generate inaccurate
estimates of genotypic frequencies and inflated Fts values. With regard to this
study, significant FIs values were initially observed at A02, HBDH, MP1 and
PEP-LA. Because scoring inconsistencies could not be eliminated as the basis
for the significant Fls values at AC-2 and MP1, these loci were excluded from
the F-statistic analysis. Scoring inconsistencies were not a problem with regard

to HBDH and PEP-1A.

Differential selection vs. gene flow
For the 8 allozyme loci exhibiting neutral variation, the absence of

substructuring among populations is due either to the mitigating effects of gene

28
flow or to very large effective population size. Mark-recapture studies suggest
that P. glaucus butterflies are highly mobile (Lederhouse 1982 and R. C.
Lederhouse unpubl., J. L. Bossart unpubl.) and capable of sustained flight.
Suitable habitat occurs continuously throughout much of the butterfly’s range
and major ecological constraints seem unlikely. Even in southern Florida where ,
suitable habitat is patchy and P. glaucus populations are reduced (Lederhouse
and Scriber 1987, Lederhouse personal communication), gene flow probably
occurs frequently. Adequate hosts and nectar sources are scattered throughout
most residential areas and around the numerous small lakes, providing sufficient
opportunity for butterfly movement among habitat patches. Since only a few
migrants are necessary to prevent random drift regardless of population size
(Slatkin 1987), we suggest that gene flow resulting from migration among
populations is responsible for the lack of differentiation among populations at
loci harboring neutral variation. Our estimates of F51, which are in the range
generally observed for other relatively mobile insects (see McCauly and Eanes
1987), further support this interpretation.

Extensive gene flow in P. glaucus would necessitate strong selection to
permit differentation in oviposition preference and larval physiology. However,
the fitness differences documented in this study appear to be more subtle.
Mortality rates were relatively equal across hosts and differences in fitness were
due to differences in relative rates of growth. Such differences would likely
translate into an increased probablity of death through attack by predators or
parasitoids (Scriber et al., unpublished), but would probably not result in

absolute mortality of nonadapted genotypes. A more likely interpretation is that

29
gene flow in P. glaucus is sufficiently restricted among populations such that
more subtle selective factors are able to operate. Direct observations of
P. glaucus movement in the field suggest that genetic variation in willingness to
leave an area, habitat quality and habitat distribution are each important in
determining how readily butterflies move about (R. C. Lederhouse unpubl., J. L.
Bossart unpubl.). We hypothesize that such factors tend to restrict migration
and gene flow in P. glaucus, thereby enabling selection to operate at a more
moderate level.

The presence of negative genetic correlations across environments is
required for genetic variation to be maintained if gene flow is equal between
environments. In their absence, optimization across hosts will eventually evolve
such that genotypes with high fitness on all hosts will predominate. In light of
this prediction, the apparent absence of tradeoffs in performance of P. glaucus
on different hosts and lack of optimization across hosts must be addressed.
Selection on Florida P. glaucus for enhanced use of M. virginiana did not appear
to be associated with a decreased ability to use either L. tulipfera or P. serotina.
Moreover, despite the fact that at least 20,000 generations have elapsed since
glacial retreat during the Pleistocene, the Ohio population in particular
continues to harbor a high frequency of genotypes not adapted to M. virginiana
(see also Bossart, this volume). We offer three explanations for these
contradictory patterns. First, it may be the case that host tradeoffs do exist but
are associated with hosts not examined here or are unrelated to nutritional
physiology per se (e.g., the ability to tolerate or minimize attack by natural

enemies associated with different hosts). Second, there may not be a single,

30
optimum genotype, such that a state of equilibrium is never obtained. Rather,
the optimum genotype may be dynamic, reflecting a stochastic, constantly
changing environment. Third, gene flow may not be balanced between
environments.

The spatial differentiation in oviposition preference and larval
nutritional physiology in P. glaucus contrasts with the lack of any such patterns
for electrophoretically detectable variation. P. glaucus behaves as one large
population with regard to neutral variation, but not with regard to ecologically
important variation. These contrasting patterns imply that differential selection
among populations (also observed within populations; Bossart, this volume) for
ecologically important traits will be counteracted by gene flow, thereby
maintaining genetic variation for host use traits. Key to this argument are
environmental heterogeneity, genetic variation for traits associated with fitness,
and genotype by environment interaction (Mitchell-Olds 1992). These three
criteria are satisfied by P. glaucus. Genetic variation was documented for both
oviposition preference and larval physiology. More importantly, this variation
was associated with differential recognition of and performance on different host
species. We predict that selection on ecologically important variation in
P. glaucus will continue to be counteracted by gene flow, thereby maintaining
genetic variation for host-use traits. In the absence of more substantial barriers
to gene flow, P. glaucus will continue to be comprised of a mosaic of genotypes

exhibiting differing abilities to use different host species.

MANUSCRIPT 2
Genetic and nongenetic components of

oviposition preference variation within
tiger swallowtail butterfly populations

Abstract

Oviposition preference might be regarded as the major determinant of
resource use for insect species which spend the duration of their larval period on
the host selected by the ovipositing female. Yet, the extent of genetic variation
within local populations, the factors affecting the phenotypic expression of host
preference, and the adaptive significance of genetic variation remain relatively
undocumented. In Papilio glaucus, the expression of preference is a function of
both genetic and nongenetic factors. Additive genetic variation for oviposition
preference appears to be common and to reflect variation in host specificity,
rather than variation in rank order of hosts. Nongenetic factors, such as eg load,
prior larval experience, and to a lesser extent age, modify the phenotypic
expression of preference. This contribution of nongenetic factors is not sufficient
to prevent selection from effecting change however. P. glaucus females exhibiting
L. tulipifera preference produced progeny that grew at a faster rate than progeny

of females that did not prefer L. tulipifera.

31

 

32
Introduction

The evolution of resource use in a phytophagous insect is a function of
the insect’s ability to behaviorally recognize and physiologically use a particular
host plant. These two traits may not share equal evolutionary importance,
however. For those species which spend the duration of their larval period on
the host selected by the ovipositing female, oviposition preference might be
regarded as the major determinant of resource use, and larval tolerance as the
more secondary factor. For these species, variation in oviposition preference (or
the lack thereof) will set constraints on future evolutionary trajectories and will
determine the direction of these trajectories. Even though larvae might be
capable of adapting to a broader array of hosts (Wildund 1973, 1975), variation
for larval tolerance of novel hosts will remain effectively neutral if ovipositing
females do not select these hosts.

Despite the significance of oviposition behavior in determining resource
use, our knowledge regarding the evolution of host preference is fairly
rudimentary in comparison to that of physiological adaptation. A number of
basic issues remain inadequately explored. Determining the degree of genetic
variation for oviposition preference within local populations is a fundamental
necessity, since it is intrapopulation variation that delineates evolutionary events.
While such variation does not appear to be uncommon (see J aenike and Holt
1991 for refs.), the current data base needs to be increased substantially before

generalizations regarding the extent of genetic variation in host preference will be

possible.

33

A second requisite is characterizing the phenotypic expression of
intrapopulation genetic variation. Phenotypic heterogeneity in host choice among
individual females may result from genetic variation in willingness to accept less
preferred hosts (i.e., host specificity or acceptability), genetic variation in
individual preference hierarchies (i.e., the sequence in which particular hosts are
ranked by individual females), or some combination of these two factors
(definitions of acceptability and preference sensu Singer 1986). Courtney et a1.
(1989) proposed a unidimensional model of host choice. In this model, each host
encountered by a female has an intrinsic acceptability, resulting in a hierarchical
ranking of the hosts. The threshold of acceptability may change over time such
that less preferred hosts become more acceptable, but not at the expense of hosts
initially preferred. An important assumption of the unidimensional model is that
genetic variation in the hierarchical ranking of hosts by individual females is not
present, i.e., females are not "free to evolve separate affinities" for different host
species (Singer et a1. 1992). In contrast, the multidimensional model of host
choice permits genetic variation among individual females in the hierarchical
ranking of hosts. As such, the probabilities of preferring different hosts are
independent for individual females (Singer et al. 1992). Until intrapopulation
variation is characterized, it will be impossible to evaluate which of these models
more closely describes reality.

A third necessity is establishing whether observed genetic variation is
adaptively significant. J aenike (1990) argues convincingly that genetic variation
for oviposition preference is effectively neutral, with host choice being determined

most frequently by the physiological state and past experiences of ovipositing

34
females. The expression of host preference may be influenced by eg load
(Odendaal and Rausher 1990, Minkenberg et al. 1992), age (Jaenike 1990), larval
host environment (Hopkins 1917, Jaenike 1983) or previous adult experience
(Jaenike 1988). Clearly, if genetic variation is commonly masked by nongenetic
factors, then the evolution of host choice will be the result of random processes.
A mechanistic exploration of an adaptationist paradigm to explain the evolution
of oviposition preference will be of nominal benefit therefore in expanding our
understanding of host use patterns in phytophagous insect populations.

As a model system, Papilio glaucus, the eastern tiger swallowtail butterfly,
provides a unique opportunity for investigating a number of these issues.
Techniques are available for mating laboratory reared adults and assessing host
preferences. In addition, many of the biological attributes which characterize this
species are especially amenable to an investigation of evolutionary aspects of host
preference. P. glaucus is a polyphagous herbivore using hosts from at least 7
different plant families. Favorite hosts include Liriodendron tulipifera, tuliptree;
Magnolia virginiana, sweetbay; Ptelea tn'foliata, hoptree; Prunus serotina, black
cherry and Fraxinus americana, white ash. In P. glaucus, the ovipositing female
determines the host substrate of her progeny since larvae rarely switch hosts.
Females will oviposit on a number of different host species, even those that do
not support larval development (Berenbaum 1981). Larvae will generally initiate
feeding on most plants, even non hosts (Feeny 1991). No single host species is
coincident with the geographic range of P. glaucus, an area which extends south
from Michigan to Florida and west to Texas (Scriber 1983), permitting local host

abundance to effect evolutionary change. Earlier studies have described patterns

35
of differentiation in oviposition preference among P. glaucus populations based on
regional host abundance (Bossart and Scriber 1993). In this study, an analysis of
genetic and nongenetic preference variation within populations of P. glaucus is
presented. Objectives were to 1) determine the extent of phenotypic variation in
host preference for an array of hosts, 2) assess what proportion of this variation is
genetically based and adaptively significant, and 3) identify potential nongenetic

factors affecting preference variation.

Materials and Methods
General

Field collected butterflies were placed in individual glassine envelopes,
then transported on ice or shipped using overnight delivery to the laboratory at
Michigan State University, East Lansing, Michigan. Females were fed a 20%
honey solution immediately upon arrival, then kept at 24°C for 24 h. This
procedure extended the lifespan of ovipositing females.

Oviposition preferences were assessed in 2~choice laboratory trials. The
initial host comparisons used to assess phenotypic variation included Liriodendron
tulipifera, Prunus serotina, Fraxinus amen'cana, Ptelea mfoliata and Magnolia
virginiana. Subsequent genetic analyses concentrated on L. tulipifera and
M. Virginiana comparisons. The decision to focus on L. tulipifera and
M. virginiana was based on the extensive phenotypic variation observed in initial
2-choice trials and the fact that both are of the Magnoliaceae family and have
non-overlapping arange distributions; hence, were interesting hosts to compare.

L. tulipifera and P. serotina foliage was collected at least every third day from

 

 

36
various areas in the vicinity of the Michigan State University (MSU) campus and
stored at 7°C. M. virginiana foliage was collected daily from potted trees
maintained in a campus greenhouse, and originally purchased as nursery stock in
Florida. Although the same host species growing in different geographic regions
likely differ morphologically and chemically, relative differences experimentally
documented among populations should indicate real variation as long as all
sampled populations are assayed simultaneously on foliage collected from the
same sources and randomly distributed. Foliage sprigs presented to ovipositing
females were placed in water-filled, rubber-capped plastic vials to maintain leaf
freshness.

Females were individually placed in clear plastic (10 x 20 x 27 cm)
"shoeboxes" (Tristate Plastics) with sprigs of appropriate foliage and fed a 20%
honey solution daily. Boxes were maintained under artificial illumination,
alternating 4h:4h photozscotophase. Fresh host sprigs of similar leaf surface area
were positioned in opposite corners along the long side of the box towards the
lights. Females were free to move within the box and were commonly observed
fluttering between host sprigs. The position of each sprig was alternated every 24
h to control for positional effects and foliage was changed as needed. The
position of individual boxes in relation to the light source was randomized at each
feeding. The few stray eggs that were placed on the paper lining or box were
generally adjacent to a particular leaf but nonetheless excluded from the analyses.
Eggs were counted and collected daily. Preferences were calculated as the
percentage of total eggs oviposited on a given host. Since percentages based on

small sample size are not very accurate, only females ovipositing at least 20 eggs

 

 

37
were included in the analyses; the majority of females ovipositing at least some
eggs, oviposited at least 20.

Laboratory reared butterflies used in preference trials were collected as
pupae and placed in individual screen cages until adult emergence. All newly
emerged females were fed a 20% honey solution, left for 24 h at 24°C, then
stored at 8°C until being mated. Newly emerged males were stored at 18°C and
fed daily, alternating a 20% honey solution with male "elixir", an ionic solution
that increases male fertility (Lederhouse et al. 1990). Males were fed for at least
six days before being mated; females were mated as soon as possible. Butterflies

were hand-paired and females set up for oviposition.

Assessment of phenotypic variation

Adult female butterflies were collected from the north-central region of
Georgia (Clarke Co.) during August, 1988. Newly captured females were shipped
to the laboratories at Michigan State in 4 separate groups at approximately 1
week intervals. Each group of females was used to compare a different 2-choice
combination, with each trial including L. tulipifera and one of four alternate hosts:
P. serotina, M. virginiana, P. mfoliata or F. amen'cana. A total of 125 butterflies
were tested. Oviposition trials lasted for 4 days. The criterion that at least 20
eggs be oviposited was met by 21 of 22 females assessed in the L. tulipifera/
P. serotina comparison, 37 of 41 for L. tulipifera/M. virginiana, 21 of 24 for
L. tulipifera/P. trifoliata and 33 of 38 for L. tulipifera/F. americana.

 

38
Genetic component of preference variation

The heritability of preference was assessed in 1989 and again in 1991
using data from second generation females collected in late July within a 15km
radius in southern Ohio (Lawrence Co.). The same experimental protocol was
followed in both years. Relative preferences of mothers were assessed for
L. tulipifera and M. virginiana. Ten to twelve newly eclosed, full-sibling larvae
from 7-12 mothers (1991 and 1989, respectively) were reared individually to
pupation in (150 cm x 25 cm) screened, plastic petri dishes containing sprigs of
P. serotina. Larvae were reared on a common host to minimize maternal effects
and to control for potential inducible responses resulting from larval host
environment. Preferences of mothers were regressed on the average preference
of their daughters following methods of Becker (1984) when only one parent is
measured and h2 = 2 [(covxz/varxfl.

Butterflies collected from Ohio in 1989 were also used to examine the
relationship between female preference and larval performance. Eight to 30
full-sibling larvae from each of 12 mothers were randomly allocated to
L. tulipifera and M. virginiana. Larvae were reared individually to pupation on
the appropriate host foliage. Larval duration, pupal mass and sex were recorded

for each individual. Relative growth rate (RGR) was calculated as,

RGR = (WP - Wr)/((Wp - Wr/Z) x D). where
(W,) is initial larval mass,
(W) is pupal mass, and

(D) is larval duration.

 

 

39
Non-genetic component of preference variation

Female attributes potentially affecting the phenotypic expression of host
preference examined in this study include: age, pupal mass and fecundity (as
measures of adult vigor), egg load and larval host environment.

Butterfly age was estimated using wing condition class; previous studies
have shown that this measure correlates well with adult longevity (Lederhouse
1983). Wing condition class was determined for a total of 88 females collected
from Lawrence Co., Ohio in 1991. Females were ranked on a scale of 1-4
(following Lederhouse 1983), with fresh, unworn females being assigned a value
of 1 and well worn females, a value of 4. Intermediates were assigned values of
either 2 or 3. The majority of the 88 females collected were ranked as either 1
or 2, indicating that sampling must have occurred relatively soon after second
generation butterflies began to emerge and become active. Of these 88 females,
50 were set up in L. tulipifera/M. virginiana 2-choice trials and the other 38, in
M. virginiana/P. serotina trials. In each of these 2-choice trials, a more preferred
host was compared against a less preferred host. Preference rankings were based
on results from the 2-choice comparisons assessing phenotypic variation. Each
wear class was split approximately equally between the two preference trials.
Preference trials lasted for four days.

The effects of fecundity, egg load and pupal mass on preference were
measured on laboratory reared, second generation progeny of butterflies collected
from Lawrence Co., Ohio in 1990. First generation neonate larvae were reared
individually to pupation on P. serotina. Emerging adults were hand paired, and

females set up for oviposition. Subsequently, 10 to 15 second generation neonate

 

 

40

larvae from each of 12 mothers were reared individually to pupation under
diapause conditions (21° C, 12:12 photo:scotophase) on either P. serotina or

L. tulipifera. Pupae were weighed to the nearest mg 24 hours after shedding the
larval skin. Diapausing pupae were stored at 8°C. The following year, emerging
adults were hand-paired and females tested in L. tulipifera/M. virginiana 2-choice
trials. Females were allowed to oviposit until death occurred; dead females were
stored at 4° C for subsequent dissection. Fecundity was measured as the number
of eggs oviposited per day.

Generally, egg load is either measured as total eggs oviposited or as total
mature eggs produced, i.e., eggs oviposited plus the number of mature eggs
dissected from the ovaries of dead females (Minkenberg et a1 1992). An analysis
using total eggs oviposited provides little information since this estimate may or
may not correlate with actual egg load; hence, the better measure of egg load is
total mature eggs produced. Egg load was estimated both directly and indirectly
to account for an unknown egg maturation rate. Direct estimates of egg load
based on total mature eggs are valid only if the majority of eggs mature
concurrently or if maturation rate and total egg load are positively correlated.
This technique is not legitimate if maturation rate is variable or uncorrelated with
total egg load. Direct estimates of egg load were obtained by summing the
number of eggs oviposited and the number of mature eggs dissected from the
ovaries of dead females to obtain total mature eggs. This number was then
divided by pupal mass to determine the number produced per gram weight, a
technique that controls for possible correlation between mass and total egg

production. Indirect estimates were obtained by analyzing trends in preference

41

over time for each ovipositing female, permitting the detection of any increasing
or decreasing patterns of preference associated with changes in egg load from day
to day independent of maturation rate.

Effects of larval host substrate on preference were examined using two
difierent host comparisons. In one comparison, relative preferences for
L. tulipifera and P. serotina were compared for a population of females reared on
L. tulipifera and a population reared on P. serotina. In the second comparison, is
relative preferences for L. tulipifera and M. virginiana were compared for females

reared on L tulipifera and M. virginiana. Newly eclosed larvae generated from

 

females collected from Lawrence Co., Ohio were reared individually to pupation r
on the appropriate host foliage. Newly emerged adults were mated, and females

allowed to oviposit until death occurred. Effects of larval host substrate on

L. tulipifera and P. serotina preference were also assessed within a family. Ten to

15 full-sibling larvae from a single Ohio field-collected mother were randomly

allocated to either L. tulipifera or P. serotina. Newly emerged females were

mated to field collected males, then allowed to oviposit. F

Statistical analyses

All linear regression and correlation analyses were performed using SAS ' 1

 

(1985). Correlation analyses were performed on family means unless stated
otherwise. Fecundity, pupal mass and egg load were analyzed in terms of a
preference index. Preference index measures the degree of preference expressed
irrespective of host species and was calculated as the absolute value of the

difference between actual percent of eggs oviposited on L. tulipifera and 50% of

42

eggs oviposited on L. tulipifera. Preference index = 0 for females ovipositing
50% of their eggs on both hosts; preference index = 50 for females ovipositing
100% of their eggs on one particular host species. Trends in preference over
time were analyzed for individual females using the nonparametric D-statistic
(Lehmann 1975). Statistical significance for the population of females was
determined as Zi = (Di-uD,)/ai, where the sum of the Z,2 ’s are approximately
distributed as Chi-square with n degrees of freedom. Effects of larval host

environment were analyzed using Chi-square analyses on 2-way frequency data.

Results

Phenotypic variation

Georgia P. glaucus females expressed extensive phenotypic variation in
host preference in 2-choice trials, and exhibited differing degrees of relative
preference for L. tulipifera depending on the 2-choice combination offered
(Figure 4). In the L. tulipifera/P. tnfoliata comparison, neither host seemed to be
clearly preferred over the other. Forty-eight percent of the females oviposited
60% or more of their eggs on P. mfoliata, and 38% of the females oviposited
60% or more on L. tulipifera. This contrasts with the distinct preference
expressed when the choices presented were L. tulipifera and P. serotina, where
81% of the females oviposited greater than 70% of their eggs on L. tulipifera.
Females also exhibited preference for L. tulipifera (although to a lesser extent)
when offered either the L. tulipifera/F. americana comparison or the
L. tulipifera/M. virginiana comparison. Seventy percent of the females presented

with L. tulipifera and F. americana oviposited 60% or more of their eggs on

 

 

68.8ch N 8 380.2053 E 6888.... .2 6588“. .m "38: 0880.8 88 .8 08 98 8.8%... d mew—0980
£05 008:0-N E 808008 083 0.00888:— 80588 4 8.« 00:08.88 02:22 80:00 8888505
08800 85:00 £8: 8on 880:8 00—080“ .303» .m .3 808290 80:08.88 8 8:83me .0 083m

asmbafifiwd co «mac 8 Econom

43

00..

cm

cm

0%

ON

 

A_Nuzv
2338.: .m
.n>

368830 .N

 

Aunuzv
afidwfimmswa .i
.2,

3.5.830 .N

 

 

00—

cm

on

0%

ON

 

Annuzv
dgdotusd RN
. u>
885.923 .N

A_Nuzv
use...» 0.3» .m

. a)
3.5.830 .N

 

 

ow
on

0—
ON
on
0v
om

salowaj jo ruaoaed

44
L. tulipifera; over half (57%) of the females presented with L. tulipifera and
M. vuginiana oviposited 60% or more on L. mlzpzfera. When ranked relative to
L. tulipifera preference, the order of preference expressed by Georgia females for
the five hosts was L. tulipifera = P. tn'foliata > M. virginiana > F. americana >

P. serotina.

Genetic component of van'ation g.
Mother/daughter regressions: The average L. tulipifera preferences

expressed by mothers when given a choice between L. tulipifera and M. vilginiana

 

was similar to that expressed by their daughters when given a choice between the ’
same two hosts (Figure 5). In general, only two preference states were expressed:
no preference and L. tulipifera preference. None of the females tested expressed
strong M. virginiana preference. With data from both years combined, the
positive correlation between mothers and daughter-averages was readily apparent
(p<.02). Only 1 of the 16 families failed to exhibit a corresponding
mother/daughter preference. In this particular 1989 family, the mother preferred
M. virginiana, while the daughters preferred L. tulipifera. The lack of agreement

between this mother and her daughters in combination with the fact that her

 

daughters exhibited very similar levels of preference, suggests that the preference
expressed by this mother was probably not genetically based.

With data from each year viewed separately (as is more appropriate when
estimating heritabilities), the positive association between relative preferences of
mothers and daughters was significant in 1991 (p<.02), but not in 1989 (p=.35).

The lack of significance in 1989 was due to the one atypical family in which

45

.n05n88 E00» Sen Bob 83 3:85 .0800an

mm 0:: c0320 05 $30» Baggy Bob 3% 5:85 .0030ch
0.8 8:: 2:5 .80Ewswc mo 00:282.“ 30525:. .4 508

no E052: we 02.2805 E0b§é 4 no commmawom .m 8&5

0.32030“. .0 00:30.93 dbmbafifiud

oo— om ow 9v ON 0

 

 

 

 

F _ _ _ p . o 7
.1.

m.

1.

.8 M

3

I.

0

lo... m

.....

J

u

now u

w

18m

0

o o 0020 u
3

mePO 100—

 

 

 

46
daughters did not exhibit the same preference as their mother; the association
was significant (p=.03) with this family excluded from the analysis. Estimated
heritabilities were high in both years (122:.71 in 1989 and 1.13 in 1991), indicating
that expressed variation in relative preference for L. tulipifera is at least partially
genetically controlled.
Preference/perfonnance correlations: In general, the average relative

growth rate of a female’s progeny was positively correlated with her preference

 

n-

I

for L. tulipifera (Figure 6). This correlation was significant on L. tulipifera (r=.65, i
p=.02), and nearly so on M. virginiana (r=.53, p=.07). Relative growth rates of

full-siblings were positively correlated across both hosts (r=.65, p=.02), arguing ’

against ecological specialization on different hosts as the basis for the

preference/performance correlation.

Nongenetic component of variation
Female Age: Twenty-three of the 50 females included in the
L. adipifera/M vizginiana comparison, and 11 of the 38 included in the
M. virginiana/P. serotina comparison, oviposited sufficient numbers of eggs to be

included in the analysis (n=20). These numbers also reflect the fact that

 

generally only a third of field collected females will oviposit under laboratory
conditions. In both 2-choice comparisons, the relative preferences expressed by
young females (i.e., those with wing condition rankings of 1 or 2) ranged from
L. tulipifera preference to M. virginiana preference (Figure 7). Younger females
as a group exhibited all three preference states: preference for the more

preferred host of the two hosts offered, preference for the less preferred host and

47

9.059.: .6 00:00:05 6.0.0.3933“ :N

sags .=
0:0 E0b§é 4 :o .80ch :0: 00 0:0: 53on 030—0: :008 05
0:0 0:050:— 00 00:20.02: EQREE 4 :00300 5:20:00 .0 050E

 

 

 

00P00 00 0h 00 on 0: 0m 0N op 0
_ _ P _ _ _ _ _ L _ No.0
.. no.0
.. V0.0
i 00.0
' O dfidwfiemewa SN 0 fl
80.0023 .q o

 

 

 

00.0

(Aop/fi/B) AueBOJd to you

48

000.0 cozzucoc .803

.88 0058\EQ
8 808000 :8 :8: oom .000 0060—0: .8 88085 03000: 0 «.0 80880
0003 00—0800 8 0.000 80880 803 05 8 800:3 0 m0 :8: 00:80::

3.0— 0 «05000 00:00:80 :08, 38: 0008008 :8 9006000 s 0805

 

 

 

 

 

0 m. N P
_ _ _ _
dfiwuokgfi .n> 028308.88: O I Dr
0:32.093?! .m> dsobafiaud O i
1 0N
. - on
a e .
_.—.u& 0 .. 0V
30.33.39. 0 m H. on
e ..
o . 8
i on
I 00
O .
o 1 cm
0 .
o o . 00F

 

 

tsoq pauaieJd uo $663 34

49
no preference. In contrast, all 7 older females (i.e., those ranked as either a 3 or
a 4) exhibited only one preference state: a strong preference (70% or greater of
deposited eggs) for the higher ranked host of the two hosts offered. Simple
linear regressions of wing condition class on relative preferences were not
significant however, and did not support this implied pattern of increased
specificity in older females.

Female Fecundity and Papal Mass: The physiological "condition" of
P. glaucus females in terms of fecundity and pupal mass did not appear to affect
female preference. Neither of these traits correlated with preference (Figure 8).
Moreover, there were no differences in the distribution of preferences expressed
by L. tulipifera reared females and those expressed by P. serotina reared females
towards L. tulipifera and M. virginiana. This similarity is despite the fact that
L. tulipifera reared females were significantly heavier on average (F[1,101=12~53’
p=.005; Figure 8) than P. serotina reared females.

Egg Load: The average number of mature eggs produced by L. tulipifera
reared females per gram of pupal mass was significantly greater than the number
produced by P. serotina reared females (FILM: 15.47, p=.002; Figure 9). More
importantly though, egg load appeared to have a significant impact on the level of
preference expressed. Females with greater egg loads tended to be ”no
preference” females, ovipositing relatively equal numbers of eggs on each host
and resulting in low preference indices (Figure 9). Females with smaller egg
loads generally exhibited a distinct preference, ovipositing the majority of their
eggs on a single host and resulting in high preference indices (Figure 9). The

relationship was significant for P. serotina reared females; the relationship was not

50

 

80 ‘ O reared on Peer-print),
70 O reared on L.tultpifera

60- O
50:
40‘
30-
20-
10-

 

 

Eggs oviposited per day

 

 

 

 

1.3:
1.2-
1.1-

 

Pupol moss

1.0 -
0.9 ~
0.8 -

 

 

 

 

l l I l l l

O 10 20 30 4O 50
Preference index

Figure 8. Relative degree of preference exhibited by females as a
function of female fecundity and pupal mass. Females were tested in
L tulipifera/M. Virginiana 2-choice trials. Solid lines are regressed
through data from different hosts; the dotted line is regressed
through data from both hosts combined. See text for definition of
preference index.

 

200 A O L. tulipifera reared
O P.3erotina. reared

 

 

 

 

 

 

U!

(n

O

E 175

—O O

O.

3

0- 150

E

2

0 125

\

(0

m

g 100

O

t.

3 75

O

E

=0: 50-
I I I I I I
O 10 20 30 40 50

Preference index

Figure 9. Relative degree of preference exhibited by females as a
function of egg load. Females were tested in L. tulipifera/

M. virginiana 2-choice trials. Solid lines are regressed through data
from different hosts; the dotted line is regressed through data from
both hosts combined. See text for definition of preference index.

52

significant for L tulipifera reared females although a similar trend was observed.

When viewed as day to day change in individual females, egg load did not
appear to affect expressed levels of preference (X2[21]=22.1, p>.3). For most
females, relative preference remained similar from day to day. Only three of the
21 females tested exhibited a significant change in preference over time. For two
of these females, preferences declined; for the other, preference increased.

Larval Host Environment: The relative distribution of eggs on F
M. virginiana and L. tulipifera was similar for both the M. virginiana reared

population of females and the L tulipifera reared population of females

 

(X2m=l.77, p>.1). However, the relative distribution of eggs on P. serotina and E.
M. virginiana differed between P. serotina reared females and L tulipifera reared i
females (X2m=68.88, p< <.001). Three of 5 (60%) L tulipifera reared females

oviposited greater than 70% of their eggs on L tulipifera, while only 3 of 11

(27%) P. serotina reared females did (Figure 10). Within a single family,

P. serotina reared females were also more willing to oviposit on P. serotina than

their L tulipifera reared sisters (X2m=29, p<.001; Figure 11).

Discussion

Three general conclusions regarding the evolution of oviposition

 

preference in P. glaucus are evident from this study: 1) females exhibit genetic
variation for host preference, 2) nongenetic factors contribute to phenotypic
heterogeneity in oviposition preference, and 3) genetic variation for preference is

adaptively significant despite the contribution of these nongenetic factors.

53

 

100 -reared on Ltul'ip'ife'ra. (N=5)

‘ -reared on P.serot'ina (N=11)
80 -

60~

4o-

20:

 

\\
§
\
s
s
s
h
s
h
s
s
\
§
§

 

 

 

 

100 -reared on Ltulip'ifera' (N=10)

‘ -reared on Mm'irg'in'iana (N=5)
80 ~

Percent of females

60-

4o-

20-

 

WW

 

 

 

 

O I I I l l
10 20 30 40 50 60 70 80 90100

L. tulipifera preference

Figure 10. Distribution of preferences for L tulipifera expressed
by a population of P. glaucus females as a function of larval host
substrate.

54

 

    

100
-L.tulipifera reared
.. -P.serotina. reared
E 80 ~
'7)
O
.9
5 60 -
m
U)
3
_._ 4o —
c
0
O
L.
3 20 -
o m

 

 

 

 

 

P. serotina L. tulipifera

Figure 11. Mean (2 SE) preferences for L tulipifera and P. serotina
expressed by full-siblings as a function of larval host substrate.

55

Genetic factors contributing to variation in host preference: The high
heritabilities estimated in this study reflect the close resemblance between mother
and daughter preferences expressed for L tulipifera and indicate a genetic
component to preference variation. However, since only females were tested,
maternal effects can not be separated from genetic effects. If maternal effects
contribute to mother-daughter resemblance, then actual heritabilities may be
lower than the estimated values. Maternal effects are known to influence larval
performance (Rossiter 1991a, 1991b), but whether such effects influence
preferences has not been established. The use of parent-offspring regressions to
estimate heritabilities does not permit additive genetic variation (i.e., selectively
important variation) to be partitioned from other sources of genetic variation.
Hence, these estimates are of heritability in the broad sense. The similarity
among the P. glaucus estimates and those obtained for Colias eurytheme
(T abashnik et al. 1981) and Euphydryas editha (Singer et al. 1988), two other
lepidopteran species, lend confidence to these values. That genetic variation is
present in P. glaucus is not surprising given that genetic variation for oviposition
preference has been demonstrated for other phytophagous insects (reviewed by
Jaenike and Holt 1991).

When given a choice between L tulipifera and M. virginiana, the relative
preferences of Ohio daughters ranged from no preference for either host to
strong L tulipifera preference. None of the daughters tested expressed a
preference for M. virginiana, even the daughters from the M. virginiana preferring
mother. The most parsimonius conclusion is that only mothers preferring

L tulipifera and mothers with no apparent preference, were expressing a

56
genetically based phenotype. The observation that M. virginiana preference does
not have an apparent genetic basis suggests an absence of genetic variation for
the order in which different hosts are ranked by individual P. glaucus females.
Rather, variation in host preference seems to be due to genetic variation in host
specificity. Some females are inherently more willing than others to oviposit on a
less acceptable host (in this case, M. virginiana), while other females oviposit
exclusively on the more acceptable host (L tulipifera).

Whether this pattern of genetic variation for host specificity, but not rank
order, hold ups to further scrutiny remains to be confirmed. Some of the mothers
exhibiting M. virginiana preference could not be included in the analysis because
daughters either were not obtained or did not oviposit. Hence, it could not be
determined whether these particular mothers were expressing a genetically-based
preference. Many of the females tested in the other 2-choice comparisons
expressed preferences for the alternate (non-L tulipifera) hosts (Figure 4). At
least some of these females were potentially expressing a genetically based
preference; an analysis of daughters from mothers tested in one of these other
comparisons might have detected additive genetic variation for rank order. At
least one mother preferring F. americana in a separate 2-choice trial gave rise to
daughters preferring E americana (Bossart, personal observation).

Historically, genetic variation for host specificity, like that shown here,
has been more easily demonstrated than genetic variation in the hierarchical
ordering of different hosts (Wasserman 1986, Thompson 1988, J aenike and Holt
1991). This absence of variation in rank order among P. glaucus females is

predicted by the unidimensional model of host choice, whereby females do not

57

evolve separate affinities for different hosts species (Courtney et al. 1989). Other
results of this study however are more compatible with a multidimensional model
of host choice. The observation that three of the mothers tested in this study
exhibited a nongenetically based preference for M. virginiana over L tulipifera,
the more acceptable host, is counter to the unidimensional model. The
unidimensional model predicts that even when lower ranked hosts become
acceptable, they are not preferred to the exclusion of higher ranked hosts.
Hence, in a population of females we might expect to detect females that prefer
the most preferred host (in this case, L tulipifera) and "no preference" females,
but should not observe females that prefer the less preferred host (in this case,
M. virginiana). Results from the other 2-choice comparisons also contradict the
predictions of the unidimensional model. None of the females in these
comparisons should have expressed a preference for the less preferred host,
regardless of whether this variation was genetically based or not. If genetically
based, then females are exhibiting genetic variation in the hierarchical ranking of
different hosts. If not genetically based, then "motivated" females are preferring a
less acceptable host to the exclusion of a more acceptable host. Neither scenario
is consistent with a unidimensional model. Clearly, additional study will be
required before the issue of dimensionality of host preference variation can be
resolved.

Nongenetic factors contributing to variation in host preference: The presence
of substantial additive genetic variation for host preference in P. glaucus indicates
that evolutionary events, such as host shifts and host range expansion, should not

be impeded by a lack of suitable genetic variation. However, for selection to

58
effect change in patterns of host preference requires that this variation not be
negated by the contribution of nongenetic factors. Of the factors measured in
P. glaucus, both egg load and prior larval experience, and to a lesser extent age,
appear to contribute to the phenotypic expression of preference and may serve to
constrain the evolution of this trait.

Physiological factors such as eg load and age can modify the expression
of oviposition preference by altering the motivational state of the female
(Courtney et al. 1989, Jaenike and Holt 1991, Minkenberg et al. 1992).
Presumably, proprioceptors detect abdominal distension caused by increased egg
load, stimulating a higher state of motivation. Highly motivated females are
expected to exhibit reduced host specificity and oviposit more willingly on less
acceptable, lower ranked hosts (Mangel 1989, Courtney et al. 1989). Both young
females and females deprived of an oviposition site for some length of time
should have higher egg loads and therefore, higher motivational states, than older
females and females ovipositing at a normal rate (Jaenike 1990, Minkenberg et
al. 1992). The negative association between egg load and preference index
documented for P. glaucus, and positive trend between age and host specificity
support this prediction. Such relationships have also been documented for certain
dipteran and other lepidopteran species (see Minkenberg et al. 1992). In most of
these other studies however, egg load was confounded by female mass. Since
female size and number of eggs produced are generally, positively correlated (e.g.,
Blau 1981, Haukioja and Neuvonen 1985), analyzing egg load irrespective of body
size is not very useful. The egg load of large females producing a greater number

of eggs may be no different than that of small females producing a smaller

59
number of eggs. Without controlling for female body weight, the patterns
observed in these other studies are difficult to interpret, and provide little useful
insight with regard to the interpretation of the results presented here.

The lack of a time effect in day to day preferences is likely a function of
the experimental technique rather than the absence of a relationship between
preference and egg load. Only 33% of the P. glaucus females tested lived longer
than five days and over half of the females oviposited a majority of their eggs on
the first two days. Given these conditions, detecting any type of statistically
significant pattern over this short period would have been unlikely. In fact, if
total mature eggs was largely a function of the number of eggs oviposited on the
first two days, this "dumping" of eggs early in the study may explain why the direct
analysis detected a correlation between egg load and preference that was not
detectable from the indirect/time analysis.

The general health or vigor of a female might also modify the expression
of host preference by altering a female’s motivational state. Healthier, more
robust females might be expected to express a higher degree of specificity than
less robust females. Presumably, stronger, more vigorous females should be more
likely to search for an extended period for preferred hosts than their less vigorous
counterparts. When measured in terms of pupal mass and fecundity, the general
vigor of P. glaucus females did not correlate with degree of preference. Despite a
significant difference in pupal mass between L tulipifera and P. serotina reared
females, the average preference for L tulipifera and M. virginiana was no different
between these two groups. Correlations were also absent when individual females

were analyzed as separate data points.

60

The effect of larval host substrate on subsequent host preference by an
ovipositing female continues to be a t0pic of discussion (Jaenike and Holt 1991).
In part, interest in this issue has remained high because of its ecological and
evolutionary importance as a mechanism whereby learned behaviors reinforce
genetically based differences in preference. Such associations if they occur can
promote sympatric, host associated divergence (Maynard Smith 1966, Rausher
1983, Jaenike and Papaj 1992). To date, little evidence has accumulated to
support the hypothesis that larval host environment modifies ovipositional
preferences in adult females (Rausher 1983, Wcislo 1989, Jaenike 1990). Hence,
the observation that larval substrate apparently influences subsequent preferences
of P. glaucus females is especially intriguing. In general, when presented a choice
between L tulipifera and P. serotina, the distribution of preferences expressed by
P. glaucus females tends to be skewed towards L tulipifera preference (as
depicted in Figure 4), a pattern that has been observed more than once in the
laboratory. This unwillingness to oviposit more than a few eggs on P. serotina by
P. glaucus females has been observed in multi-choice host trials as well (Scriber
1993). L tulipifera reared females exhibited the beginnings of this same general
distribution of preferences skewed towards L tulipifera preference, when
presented with an L tulipifera/P. serotina 2-choice combination (Figure 10). This
pattern. was observed despite the fact that only five females were tested; three of
the five females oviposited 90% or more of their eggs on L tulipifera. In
contrast, the distribution of preferences of P. serotina reared females when
presented with this same 2-choice combination was not skewed towards

L. tulipifera preference, but rather spanned from P. serotina preferring females to

61
L tulipifera preferring females. In fact, only one of eleven females oviposited at
least 90% of her eggs on L tulipifera. In general, P. serotina reared females were
more willing to oviposit on P. serotina, a host typically not preferred. These
population distributions are based on small sample size and, as such, are not
amenable to direct statistical analysis; however, the fact that a similar link
between larval host substrate and adult females preference was observed within a
single family supports the interpretation that larval environment can modify
ovipositional preference in P. glaucus.

Only certain larval host environments appear to influence subsequent
preference in adult females. In contrast to results from the L tulipifera/

P. serotina assay, L tulipifera reared females behaved no differently than

M. virginiana reared females when presented with an L tulipifera/M. virginiana
2-choice combination (Figure 10). The similarity in preferences expressed
between these two groups of females may reflect the fact that L tulipifera and
M. virginiana are both Magnoliaceae. The suite of chemical cues perceived by
larvae and ovipositing females may be very similar for these two hosts.

Larval environment is thought to modify adult chemosensory responses in
one of two ways. The link may result because females retain some "memory" of
their larval host via an induced response. Sensory receptors may become "locked
in on” stimuli experienced during early stages of development to the exclusion of
stimuli experienced at a later stage ( =Hopkins host selection principle; Hopkins
1917). Alternatively, this link may depend not on memory, but rather on
chemical cues that are retained in the insect’s cuticle or other tissues and

perceived during subsequent stages of development (=chemical legacy hypothesis;

62
Corbet 1985). Such trace chemicals may be especially important during
susceptible periods of development, e.g., chemical stimuli incorporated into the
pupal case may be perceived by a female as she emergences. The causal
mechanism underlying the link between larval host substrate and female
preference in P. glaucus is unknown. However, determining the physiological
basis of this link is secondary to the fact that such a link apparently exists for
certain host species. The functional outcome will be the same regardless of the
mechanism involved.

Adaptive significance of host preference variation: For genetic variation in
host preference to be effectively neutral in P. glaucus would require that the
nongenetic factors modify behavior to the extent that selection is unable to
”recognize" an optimum genotype (Michaud 1990). Although nongenetic factors
do alter the behavioral responses of ovipositing P. glaucus females, this
contribution to phenotypic variation is apparently not sufficient to prevent
selection from effecting change. When tested in an L tulipifera/M. virginiana
2—choice array, P. glaucus females exhibiting L tulipifera preference produced
progeny that grew at the fastest rate on both L tulipifera and M. virginiana. Since
the preferences exhibited by these mothers had a significant genetic component
(Figure 5), the existence of these positive correlations indicates a selective
advantage. The interpretation that genetic variation is adaptively significant in
P. glaucus is consistent with results that indicate that genetic variation in
oviposition preference has lead to differentiation among geographic populations,
presumably as a result of selection for increased recognition of locally abundant

hosts (Bossart and Scriber 1993, Bossart, this volume).

63

The observation that relative preferences for L. tulipifera were positively
correlated with relative growth rate on both L tulipifera and M. virginiana
constrasts with the response typically observed. In general, correlations between
female preference and offspring performance arise because females choose the
plant species (Via 1986, Singer et al. 1988) or plant type within a species (Ng
1988) most suitable for offspring performance among the choices presented. Such
correlations are purported to be a mechanism whereby sympatric divergence can
result (Smith 1966). The results presented here differ from these other studies in
that preference for a single host (L tidipifera) correlated with performance on
multiple hosts. Since the P. glaucus females used in this experiment were
collected fi'om Ohio and never actually encounter M. virginiana, selection for
M. virginiana tolerance has not occurred. That selection has increased recognition
and tolerance of L tulipifera without a corresponding loss of M. virginiana use,
indicates the lack of a tradeoff between M. virginiana use and L tulipifera use.
Despite the fact that M. virginiana is a less suitable host in general (the fastest
growing families on M. virginiana still grew at a slower rate than the slowest
growing families on L tulipifera), the significant positive correlation between
relative grth rates of full-siblings on the two hosts suggests that larval tolerance

on L tulipifera and M. virginiana is controlled by many of the same genes.

MANUSCRIPT III
Intrapopulation genetic variation and covariation

in larval performance across hosts
in the polyphyagous eastern tiger swallowtail

Abstract

Full-sibling Papilio glaucus larvae from different geographic populations
and broods expressed differential abilities to use three different host species:
Prunus serotina, Liriodendron tulipifera and Magnolia virginiana. The significant
host x family interaction resulted from changes in rank order of family means on
the different hosts, indicating the P. glaucus is a "composite" generalist rather than
a "true" generalist. This is despite the presence of homeostatic genotypes. Host
use patterns were variable among populations and broods, arguing for the
importance of analyzing intrapopulation variation for addressing evolutionary
questions. Indirect evidence of tradeoffs in host use from multiple sources

suggests that negative genetic correlations underlie host use patterns in P. glaucus.

Introduction
Falconer (1952) viewed a character expressed in two environments as two
separate, genetically correlated traits. Such correlations among traits expressed in
different host environments are purported to be a major determinant of diet

breadth in herbivorous insects. Host generalization tends to be less common than

65
specialization (Strong et al. 1984, Bernays and Graham 1988), implying that
physiological tradeoffs in the ability to use multiple hosts dictate the evolution of
diet breadth. However, even generalists may be comprised of a mosaic of
differentially adapted genotypes exhibiting relative specialization on different
hosts (Fox and Morrow 1981).

The extent to which tradeoffs in performance underlie host use patterns
is unresolved. To date, there is little evidence to support the contention that
negative genetic correlations define degree of specialization (Rausher 1984,
Jaenike 1989, Fry 1990). To further confound the issue, Fry (1993) demonstrated
that tradeoffs can exist even when genetic correlations based on full-sibling family
means are zero or positive, and may not be present even when genetic
correlations based on family means are negative. These spurious statistical results
led Fry (1993) to conclude that research efforts to investigate correlated
responses should focus on selection experiments. Unfortunately, the use of such
experiments to resolve whether tradeoffs occur will likely not prove applicable in
many cases. For many ecologically interesting insect species, selection
experiments are so labor intensive as to be effectively impossible to carry out.
Moreover, many species are difficult to maintain over multiple generations in a
laboratory setting.

Given the problems associated not only with statistical procedures for
analyzing correlated responses, but also with selection experiments, determining
whether tradeoffs are important components of host breadth will necessitate the
use of an array of techniques. Combining information obtained from a variety of

sources should increase our interpretative abilities. When used in combination

a"

-c_

 

66
with statistical analyses, ”norms of reaction" or "the profile of phenotypes across
hosts produced by a given genotype..." (Via and Lande 1985), may be especially
useful for investigating the genotypic structure of populations. Norms of reaction
are pictoral representations of genotype by environment interactions and genetic
correlations and permit direct observation of the response of a population of
genotypes towards multiple host environments. In the absence of tradeoffs, the
rank order of phenotypes on each host is maintained (Figure 12a). When
tradeoffs are present, the rank order of phenotypes across hosts changes (Figure
12b). The degree of crossing among lines connecting family means on each host
is an indication of the extent to which different genotypes are optimum on
different hosts and the extent to which tradeoffs may be present.

Papilio glaucus L., the eastern tiger swallowtail butterfly, provides a
unique opportunity for examining the role of differential performance and
genotypic tradeoffs in defining patterns of host use. P. glaucus is an especially
polyphagous tree-feeding insect, using hosts from at least seven plant families
(Bossart and Scriber 1993). The ability of P. glaucus to feed on multiple hosts
could result as a function of a single, especially robust genotype or conversely, as
a function of differentially adapted genotypes (Fox and Morrow 1981). In this
study I examine full-sibling performance of Ohio, Georgia and Florida P. glaucus
populations on three hosts. Extensive variation among families within
populations, coupled with interactions among families and host species, would be
evidence that P. glaucus is comprised of a mosaic of relatively oligophagous

genotypes. The degree to which physiological tradeoffs underlie this variation

67

 

 

 

 

 

 

 

 

 

 

 

 

 

(D
m L
Q) : L? I“.
L I
U)
O b.
0)
L
U
C ‘ 4
‘v
‘7
7/
I I
Host 1 Host 2

 

Figure 12. Reaction norms depicting two types of patterns that
can underlie significant genotype x environment interactions: a)
the rank order of phenotypes on each host is the same, tradeoffs
are absent, b) the rank order of phenotypes is different on each
host, tradeoffs may be present.

68
would be directly linked to the extent of change in rank order performance

among P. glaucus families across different plant species.

Materials and Methods
Larval sources
Larvae were generated from adult P. glaucus females collected from field
populations in Lawrence Co., Ohio, Clarke Co., Georgia and Highlands Co.,

Florida. Butterflies were collected from several local sites in each region.

emnfifi

Florida butterflies were collected once, during early April 1988. The Ohio and

 

Georgia populations were sampled twice for comparative purposes. Butterflies

1

were collected from Ohio in July 1988 and 1991 to provide information on larval
performance across years; butterflies were collected from Georgia in August 1988
and April 1989 to provide information on larval performance across broods.
Field collected butterflies were placed in individual glassine envelopes, then
transported to Michigan State University. Females were fed a 20% honey
solution upon arrival, then stored at 24°C for 24 h prior to being set up for
oviposition. To induce oviposition, females were individually placed in (10 cm x
20 cm x 27 cm) clear plastic "shoeboxes" (Tristate Plastics) with sprigs of

L tulipifera, a preferred host of ovipositing P. glaucus females. Boxes were

 

maintained under artificial illumination, alternating 4h:4h photo:scotophase.
Ovipositing females were fed daily. Eggs were collected and stored at 24°C until

eclosion.

69
Larval performance

Full-sibling larval performance was compared across three P. glaucus
hosts: Liriodendron tulipifera, Magnolia virginiana and Prunus serotina. The
frequency of use of these hosts in nature differs for different P. glaucus
populations. M. virginiana is the only host common to Highlands Co., Florida,
but occurs only rarely in Clarke Co., Georgia, and not at all in Lawrence Co.,
Ohio. P. serotina and L tulipifera both occur commonly in Georgia and Ohio.

P. serotina and L tulipifera foliage was collected at least every third day from
various areas in the vicinity of the Michigan State University campus and stored
at 7°C. M. virginiana foliage was collected daily from potted trees maintained on
campus.

Eggs were checked for eclosion every 2-3 h from 7:30 am - 7 pm. Larvae
eclosing between 7 pm - 7:30 am were excluded from the performance assays.
Ten to 15 neonate larvae from 8 - 17 families were randomly allocated to'each of
the three host species. Larvae were reared individually to pupation in (150 cm x
25 cm) screened, plastic petri dishes containing sprigs of the appropriate host
foliage. Foliage sprigs were placed in water-filled, rubber-capped plastic vials to
maintain leaf freshness. Petri dishes were stacked randomly in grth chambers
maintained at 24°C, 18h:6h photo:scotophase. Larvae were checked daily and
flesh foliage provided at least every other day. larval duration, pupal mass and
sex were recorded for each individual. Larval duration was defined as the period
from day of eclosion to the prepupal stage, whereupon larvae cease feeding, void
gut contents and undergo a conspicuous color change from green to brown.

Pupae were collected and weighed 24 h after shedding their larval exoskeleton.

.11 __

70
larval survival was calculated as a family percentage. Relative growth rate

(RGR) was calculated as,

RGR = (W, - W,)/((WP - W,/2) x D), where
(W,) is initial larval mass,
(WP) is pupal mass, and

(D) is larval duration

Statistical analyses

Larval duration, pupal mass and relative growth rate were analyzed using
the general linear method, mixed-model analysis of variance (PROC GLM; SAS
1985). Family and family x host were designated as random effects. Expected
mean squares and error terms used to test each effect followed Ayres and
Thomas (1990). All population-collection date combinations were analyzed
separately. Standard product-moment genetic correlations between family means
were calculated for each host pair for each larval trait examined (Via 1984).
Survival data was arcsine transformed, as recommended when percentage data
spans a wide range.

Pupal mass and larval duration tend to be sexually dimorphic in
P. glaucus. When mean male and female pupal mass and larval duration were
compared for all families producing both sexes, females on average tended to be
100 mg heavier and grew 1.2 d longer than males. Not all families were
represented by both males and females, consequently, sex could not be included

as a source of variation in the analysis of variance. To account for the potential

72

Table 4. Analysis of variance comparing larval duration for full-sibling
P. glaucus families collected from different populations and broods and reared on
P. serotina, M. virginiana or L tulipifera.

 

 

Source of

Variation df MS F“
A. Florida - 1988

Host 2 36.11 7.25 “

Family 7 6.16 2.28‘

Host x Family 14 4.98 185'

Error 60 2.70

B. Georgia August Brood - 1988

Host 2 230.07 20.09""
Family 16 22.97 2.96‘"
Host x Family 32 11.45 1.48‘
Error 175 7.75

C. Georgia April Brood - 1989

Host 2 24630 18.08” ‘
Family 1 1 10.26 1.25
Host x Family 21 13.62 167‘
Error 181 8.18

D. Ohio July Brood - 1988

H08! 2 180.11 1951'”
Family 9 4.06 0.54
Host x Family 18 9.23 1.22
Error 100 7.55

E. Ohio July Brood - 1991

Host 2 852.82 118.38"‘
Family 13 14.38 2.00'
Host x Family 23 7.17 0.99
Error 141 7.20

 

‘ P < .05, ” P < .01, ”‘ P < .001.
' The F test denominator for Host was MSHO‘, x “may. The F test denominator for
Host x Family was MS

61101"

 

marl

73

Table 5. Analysis of variance comparing relative grth rate for full-sibling
P. glaucus families collected from different populations and broods and reared on
P. serotina, M. virginiana or L tulipifera.

 

Source of
Variation df MS F‘

 

A Florida - 1988

Host 2 3.0 x 10“ 6.84"
Family 7 5.6 x 10“ 2.50.
Host x Family 14 4.4 x 105 199*
Error 83 2.2 x 10"

B. Georgia August Brood - 1988

Host 2 1.8 x 10'3 28.77"”
Family 16 1.4 x 10“ 3.14m
Host x Family 32 6.4 x 105 1.40
Error 175 4.6 x 10:5

C. Georgia April Brood - 1989

Host 2 2.0 x 10'3 30.27m
Family 11 8.7 x 10" 1.30
Host x Family 21 1.4 x 10" 2.02"
Error 179 6.7 x 10.5

D. Ohio July Brood - 1988

Host 2 1.1x 10'3 22.31m
Family 9 2.8 x 10:5 0.77
Host x Family 18 4.8 x 10.5 132

E. Ohio July Brood - 1991

Host 2 3.0 x 103 127.2 m
Family 13 5.1 x 10'5 2.15"
Host x Family 23 2.4 x 10:5 1.00
Error 141 2.4 x 10"

 

‘ F< .05, " P < .01, ‘" P < .001.
' The f test denominator for Host was MSH“, x Fanny. The F test denominator for
Host x Family was MScmr

mm

74

Table 6. Analysis of variance comparing pupal mass for full-sibling P. glaucus
families collected from different populations and broods and reared on
P. serotina, M. virginiana or L tulipifera.

 

Source of
Variation df MS I“

 

A. Florida - 1988

Host 2 0.435 551‘
Family 7 0.096 2.67'
Host x Family 14 0.079 1.21
Error 60 0.036

B. Georgia August Brood - 1988
Host 2 0.339 11.69'“
Family 16 0.106 4.69‘”
Host x Family 32 0.029 1.29
Error 175 0.227

C. Georgia April Brood - 1989
Host 2 0.195 1087‘”
Family 11 0.098 5.43‘"
Host x Family 21 0.023 1.29
Error 179 0.018

D. Ohio July Brood - 1988
Host 2 0.294 8.65“
Family 9 0.114 6.46'”
Host x Family 18 0.034 1.91‘
Error 99 0.018

E. Ohio July Brood - 1991
Host 2 0.213 9.69'"
Family 13 0.060 2.73“
Host x Family 23 0.015 0.70
Error 141 0.022

 

‘ P < .05, ” P < .01, ‘" P < .001.
' The F test denominator for Host was MSH“, ,, Family. The F test denominator for
Host x Family was MSem,

75
this variance attributable to families consists of both genetic and nongenetic
maternal effects.

The most significant result with regard to differential adaptation and
tradeoffs was that P. glaucus families from Florida and Georgia responded
differently to the three host species. A significant host x family interaction effect
was present for larval duration (Table 4) and relative growth rate (T able 5) in
both the Florida and Georgia April broods. A significant interaction effect was
also present for larval duration (but not relative growth rate, p = .09) in the
Georgia August brood. Significant interactions among families and hosts were
absent in both Ohio broods for these two traits. However, the Ohio brood
collected in 1988 was the only brood to exhibit a significant interaction among
families and host species for pupal mass (Table 6).

A significant host x family interaction can be present even if there is no
change in the rank order of family means across hosts, i.e., the same families are
optimum across species. However, tradeoffs in performance are implicated only
when the host x family interaction reflects changes in rank order of family means.
The norm of reaction diagrams for relative growth rate and larval duration
indicate that P. glaucus families do not maintain the same level of performance
across hosts (Figures 13-15) and rank order does change; only relative growth
rates are depicted since this variable incorporates both larval duration and pupal
mass. For example, the three Florida brood families growing at the slowest rate
on M. virginiana, developed at the fastest rates on L tulipifera (Figure 13).
Similarly, many of the Georgia April brood families growing at the slowest rates

on M. virginiana, developed at the fastest rates on P. serotina (Figure 14). Even

76

 

0-10 ‘ Florida April Brood - 1988

0.09 -

0.08 ~

0.07 -

RGR (mg x g—‘ x day”)
l
x

0.06 -

 

 

0.05 a

 

I I I
P. serotina M. virginiana L. tulipifera

Figure 13. Reaction norms of relative grth rate for full-sibling, Florida
P. glaucus larvae developing on P. serotina, M. virginiana or L tulipifera.

-1 -1
RGR (mg x g x day)

0.10

0.09

0.08

0.07

0.06

0.05

0.10

0.09

0.08

0.07

0.06

0.05

77

 

 

Georgia August Brood - 1988

 

 

 

 

Georgia April Brood - 1989

 

 

 

P. se'lrot'ina. M. virg

iniana. L. tulipifera.

Figure 14. Reaction norms of relative grth rate for full-sibling, Georgia
P. glaucus larvae developing on P. serotina, M. virginiana or L tulipifera.
The use of the symbol indicates a family not represented on L mlrpiferu.

-1 ..
RGR (mg x g x day 1)

0.10

0.09

0.08

0.07

0.06

0.05

0.10

0.09

0.08

0.07

0.06

0.05

78

 

" Ohio July Brood - 1988

 

 

 

 

 

Ohio July Brood - 1991

‘

\“‘/

‘

 

‘ \ / //

/

 

 

I I
P. serotina M. virginiana

L. tulilpifera.

Figure 15. Reaction norms of relative growth rate for full-sibling, Ohio P.
glaucus larvae developing on P. serotina, M. virginiana or L tulipifera. The
use of the symbols indicates families not represented on M. virginiana.

79
when host x family effects were not significant and the majority of families
maintained the same level of performance across hosts, families with low rankings
on one host and higher rankings on other hosts were present, as was the case in
the Georgia August brood (Figure 14) and Ohio July brood (Figure 15).

Of the traits examined, family survival was the most labile (Figures
16-18). The Florida brood exhibited the least variation; variation was lower not
only on a given host, but also across hosts. The Ohio 1991 brood was the only
brood to exhibit an obvious host effect, with survival on L tulipifera being 40 and
50% higher on average than survival on P. serotina and M. virginiana, respectively.
However, survival of the Ohio 1991 brood on P. serotina and M. virginiana was
much lower than for any of the other broods.

There were no major differences in host use patterns between the 1988
and 1991 Ohio broods. Although the Ohio 1988 brood grew 20% faster overall
than the Ohio 1991 brood, the majority of families expressed the same general
pattern across hosts (Figure 15). It’s noteworthy, however, that two of the Ohio
1988 families were unique in their ability to grow at a fast rate on M. virginiana.
A significant difference did occur between the late season, 1988 Georgia brood
and the early season, 1989 Georgia brood. Despite virtually identical
development for both Georgia broods on M. virginiana and L tulipifera,
development of the Georgia April brood was 15% faster than the August brood
on P. serotina (25.4 d vs. 30.1 (1; Figure 14).

Pupal mass (Table 7) and survival (Table 8) were the only traits for
which significant family-mean genetic correlations were present between host

pairs. All significant correlations were positive. The particular correlated host

80

 

 

 

 

 

 

 

 

 

100 q \

3 80 _, .\ -
.2
E .
3 60 a
in

.1
'E
0 40 0
0
0 .
0
n. 20 -

O - Florida April Brood - 1988
I I I
P. serotina M. virginiana L. tulipifera

Figure 16. Reaction norms of percent survival of full-sibling, Florida
P. glaucus larvae developing on P. serotina, M. virginiana or L tulipifera.

Percent Survival

100

80

60

40

20

100

80

60

40

20

 

 

 

 

Georgia August Brood - 1988

 

 

 

 

 

 

‘ -
\
/’ v< «a

 

it
t
\‘i

\

(t

4

’

 

 

Georgia April Brood - 1
I

P. sero tina. M. virginiana L. tulipifera

Figure 17. Reaction norms of percent survival of full-sibling, Georgia
P. glaucus larvae developing on P. serotina, M. virginiana or
L tulipifera.

82

 

 

 

 

 

 

 

 

 

 

 

Percent Survival

 

 

 

 

100 -
80 -
60 -
40 - r
20 - :'
‘ Ohio July Brood - 1988
O I I I J
100 -
80 -
60 -
. 3\
4o 0 \2
20 ‘ V
- \‘x
O " Ohio July Brood - 1991

 

I I I
P. serotina M. virginiana L. tulipifera

 

Figure 18. Reaction norms of percent survival of full-sibling, Ohio
P. glaucus larvae developing on P. serotina, M. virginiana or
L tulipifera.

83

 

 

 

 

00...... .15 n... i. i- .ii

48. v "No: as. Viflt W0. V F»
80 8.: 0.2,: 80 8.: 03>: :00 8.: 00>:
$0 50 8.: 0.0.: 0:0. 80 8.: 030.: 80 0:0 8.: .08.:
0:10 :8: - 88: .203. 88 .m
80. 8.: 0.>2 .80 8.: 03>: 80. 8.: 00>:
80 0:0 8.: 0.8 4:0 80 8.: 030.: R0 :0. 8.: .08.:
0:10 80: - 08.: 8.03.. oao .:
$0 8.: 0.2.: ~00 8.: 03>: $0 8.: 00>:
0:0. 0:0. 8.: 0.8 :80 4:0 8.: 0%.: :0. 80. 8.: :08:
:0: us 80: - 08.0 0.04. 00.80 .0
0:0 8.: 052 :80 8.: 03>: ~00 8.: 00>:
0:0 80 8.: 0.8 .80 :80 8.: 030.: 80 80 8.: .08.:
G: "5 80: - 08.0 .003. 00.80 .0
0:0. 8.: 052 00 8.: 03>: 5.? 8.: 00>:
$0 80 8.: 0.8 2.? :80 8.: 038 80 80 8.: 0.8.:
8:0 80: - 08.0 =0... 02.6: a.

0.5 0.3.: 0.8 03.5 03>: 030.: 5.5 00>: .80.:

 

59000000: .Eosfii: 4 0:0 080.8: .3 65:20:. .& :0 0.0.: 0:320

300—0.: 0 8:.5 0:0 6.52 :0qu 50.500000: .8083 4 0:0 83%.: .3: .0890... .m :o 000:: .09.: n :35..— 0:0 83>: 40300 50.60090:
.88.... 4 05 33.50.: .z .82... .0 8 802:. at: u 5.5 05 00>: .58 0888.88 .8: .888 a: 8.8. 080...

-8 .3 08an 0:3. 9... .8: :25: :o o... 8802.8 333-888.: 08.: 08 250.008 0030 .0 .e: :08. 832.8 0 uses

Table 8. Correlation matrix of full-sibling larval survival across 3 host species
for P. glaucus populations and broods. PSsrv, MVsrv and LTsrv = survival on
P. serotina, M. virginiana and L. tulipifera, respectively.

 

 

PSsrv MVsrv LTsrv
A. Florida April Brood - 1988 (n=8)
PSsrv 1.00 0.40 -0.31
MVsrv 1.00 0.1.1
B. Georgia August Brood - 1988 (n= 16)
PSsrv 1.00 0.21 0.81’“
MVsrv 1.00 0.36
C. Georgia April Brood - 1989 (n=11)
PSsrv 1.00 0.04 0.30
MVsrv 1.00 0.13
D. Ohio July Brood - 1988 (n=11)
PSsrv 1.00 0.05 00.05
MVsrv 1.00 0.08
E. Ohio July Brood - 1991 (n=14)
PSsrv 1.00 0.51‘ 0.25
MVsrv 1.00 0.08

 

‘ F< .05, ”‘ P < .001.

85
combinations were brood dependent. The exception was the Georgia August
brood, where pupal mass was significantly correlated for all host combinations,
i.e., families with the heaviest mass on one host tended to attain the heaviest
mass on the other hosts (Table 7). Many of the nonsignificant host/trait
correlations were fairly large, suggesting that at least some of the same genes
were controlling development. Twice as many of the nonsignificant correlations

were positive, as negative.

Discussion

The significant host x family interactions and change in rank order of
performance across hosts provides evidence that P. glaucus is comprised of a
mosaic of differentially adapted genotypes, exhibiting differing abilities to use
different plant species. The presence of genetically based variation in diet
breadth establishes that P. glaucus is a composite generalist rather than a true
generalist (Fox and Morrow 1981), i.e., polyphagy in P. glaucus is a function of
the genotypic structure of the species as a whole, not of one individual, especially
robust genotype. However, there is also a homeostatic component of polyphagy
in P. glaucus. A number of families in both Georgia broods, but particularly the
Georgia August brood, exhibited comparable relative growth rates across all three
hosts. That these genotypes were observed only in the Georgia population is
especially noteworthy since this population is the only population in which all
three hosts occur sympatrically. The Florida population encounters only
M. virginiana and the Ohio population, only P. serotina and L. tulipifera (Bossart,

this volume; Scriber 1986). Such developmental stability is predicted as an end

86

result of differential selection in a heterogeneous, stable environment when
negative genetic correlations are absent and gene flow is balanced (Via and
Lande 1985). However, equilibrium conditions are likely extremely rare in
natural populations and since all the homeostatic genotypes exhibited lower
growth rates than any of the other families on L. tulipifera, and lower than most
other families on the other two hosts, it is unlikely that their frequency in the
population will increase unless they display an advantage on hosts not examined.

Differences in host use patterns among Georgia broods likely reflect
phenological variation in the quality of P. serotina as a food resource. A higher
nutritive value of early season P. serotina foliage would explain the faster relative
growth rates of early brood larvae. Activity levels of B-glucosidase in P. glaucus
larvae developing on P. serotina are known to be seasonally variable (Lindroth
1988), presumably reflecting seasonal change in plant glycosides. Plants, in
general, express seasonal changes in both the quality and quantity of nutrients
and allelochemicals (Mattson 1980, Scriber and Slansky 1981). That different
P. glaucus populations and broods exhibit variable host use patterns argues the
importance of analyzing intrapopulation variation when addressing evolutionary
questions. The variation among P. glaucus broods and populations illustrates that
interpretations based on population means will not adequately describe the
evolutionary potential of separate populations (Via 1990). Moreover, since the
sign and magnitude of genetic correlations are dependent upon, not only
environmental effects, but also deve10pmental stage, and specific traits examined
(Stearns et al. 1991), it may not be valid to extrapolate from population to

population or year to year.

87 ,

The degree of correlation between performance on different host species
within a given population or brood reflects the extent to which host use traits are
governed by the same alleles. Significant correlation indicates that traits are
linked or controlled by pleiotropy, while low correlation indicates that different
genes underly each trait (Falconer 1981). Most genetic correlations calculated for
P. glaucus were positive and nonsignificant, suggesting that adaptation to different
host species is generally unconstrained by other hosts. Few negative correlations
were observed and those that were, were weak and nonsignificant. It is
interesting, though, that when negative correlations were detected, they tended to
be associated with populations and hosts that might be expected to show evidence
of tradeoffs if such existed. In both the Florida and Ohio 1988 populations, larval
duration and relative growth rate on L. tulipifera tended to negatively covary with
these same two traits on M. virginiana. If tradeoffs between use of these two
hosts are present, then selection for enhanced use of locally occurring hosts
(M. virginiana in Florida and L. tulipzfera in Ohio) would decrease use of the host
not present (L. tulipifera in Florida and M. virginiana in Ohio).

The reaction norms suggest that tradeoffs in host use may be present, as
indicated by the crossing among family lines on different hosts. In all P. glaucus
broods, there were genotypes that expressed optimum performance on one host,
while expressing suboptimum performance on a second host. Tradeoffs were also
implied by the absence of an optimal genotype across all hosts; this is despite
ample opportunity for the evolution of optimization. At least 20,000 generations
have elapsed since glacial retreat, yet a predominance of genotypes exhibiting

high fitness on all hosts is absent; P. glaucus populations continue to be

88
comprised of a number of presumably nonadapted genotypes.

The evidence for genetic tradeoffs in the response of P. glaucus larvae
towards different host species is indirect, but comes from multiple sources: the
presence of significant family x host interactions of the crossing type, the presence
of weak, negative genetic correlations between traits that would be expected to
show such correlations if tradeoffs exist, and the lack of optimization across hosts
despite ample evolutionary opportunity. Even such minimal evidence necessitates
that the potential impact of tradeoffs on evolutionary trajectories be considered
(Via 1990). Tradeoffs are implicit in theories pertaining to the evolution of
specialization in polyphagous insects and can maintain genetically-based variation
in patterns of host use (e.g., Rose 1983). Differentiation already occurs between
Florida and Ohio P. glaucus for host use traits (Bossart, this volume, Scriber
1986). The presence of a tradeoff in the use of M. virginiana and L. tulipifera
could effect continued divergence between these two populations until such time
that they become completely, independently evolving entities. At the very least,
such tradeoffs could impose significant evolutionary constraints on host
adaptation, especially initially. Tradeoffs in host use can impede not only the
rate, but direction of evolution (Lande 1984, Via 1984, Arnold 1987).

The lack of convincing evidence of genetic tradeoffs from natural
populations to date is not surprising. Negative correlations should only be
apparent in populations currently in a state of evolutionary flux. If negative
correlations were the driving force underlying specialization on a particular host,
alleles permitting use of a second host would have been eliminated from the

population and hence, would not be detected. Moreover, given that selective

89
forces acting on populations are numerous and variable over time, tradeoffs in
feeding ability on different plant species would likely be concealed unless they
were particularly strong. The sample sizes needed to detect such correlations
would need to be large (Stearns et al. 1991). When coupled with the statistical
and logistical difficulties associated with testing and detecting genetic tradeoffs, it
is little wonder that their presence has gone virtually undetected in natural

populations. The approach used herein was to combine information obtained

 

w
from analyses of variance, product-moment correlations and reaction norms. l 5
Unfortunately, despite providing useful information, none of these techniques

permit unequivocal conclusions to be drawn (Fry 1993, Rausher 1983). While

laboratory selection experiments permit valid conclusions, they are impractical for a

most ecologically interesting species, and may not be representative of natural
populations since natural populations never encounter laboratory conditions.
Given the difficulties associated with demonstrating tradeoffs in host use, clear
evidence for such tradeoffs in herbivore populations will likely continue to be

elusive.

 

GENERAL CONCLUSIONS

With regard to the traits studied in this project, host adaptation in
P. glaucus is a function of both female host preference and larval physiology.
That the host environment encountered by P. glaucus is heterogeneous is
evidenced by the differential performance of populations and genotypes within
populations on different host species. Presumably, these differences in host use
are due to differences in the physical and chemical attributes of the plants.
L. tulipifera appears to be the most Optimal host species. Not only does this host
tend to elicit the highest response in ovipositing females, it also tends to support
the fastest growing larvae which develop into the heaviest pupae. Much of the
extensive phenotypic variation that exists in P. glaucus for host preference and
larval tolerance (Ayres et a1. 1991; Bossart, this volume) is genetically based. The
presence of adaptively significant variation both within and among geographic
populations provides ample opportunity for local selective factors to shape host-
use patterns. No one host is coincident with the range of this butterfly and
separate populations have differentiated with respect to their ability to recognize
and develop on locally abundant hosts. Florida P. glaucus oviposit more willingly,
and develop at a faster rate, on M. virginiana, the prevalent Florida host, than
Ohio or Georgia P. glaucus. Presumably, this response is a result of the selective

pressures imposed by this host.

90

91

The presence of significant interactions between host species and families
establishes that polyphagy in P. glaucus is not a function of a single, particularly
robust genotype, but of a collection of differentially adapted genotypes. Not all
genotypes responded equally well in all host environments; optimum genotypes on
one host were often suboptimumal on another. Evidence from indirect sources
suggests that physiological tradeoffs in the ability to use M. virginiana and
L. tulipifera, in part, underlie this differential performance. Definitive evidence,
however, will likely remain elusive. The use of selection experiments to address
the existence of tradeoffs in P. glaucus was unproductive in this study and will
probably be unproductive in ensuing studies since generations subsequent to the
F2 are extremely difficult to obtain.

The geographic differentiation observed for larval tolerance and host
preference contrasted with the lack of any such pattern for electrophoretically
detectable, presumably neutral variation. In the absence of knowledge regarding
differentiation in host use traits, the electrophoretic results indicate virtual
panmixia in P. glaucus. That P. glaucus is a single, panmictic population,
however, is extremely unlikely. Isolation by distance should preclude completely
random mating. Mating would also be nonrandom if restrictions to gene flow
prove to be present, as suggested by other studies (R. C. Lederhouse, unpubl.;

J. L Bossart, unpubl.). A more likely explanation for the absence of
differentiation among p0pulations in neutral variation is that the electrophoretic
analysis of isozymes was probably not sufficiently sensitive to detect genetic
difierentation within P. glaucus, differentiation that might be revealed by other,

more sensitive molecular techniques. This interpretation is supported by results

92

of other studies which demonstrate that an analysis of nuclear or mitochondrial
DNA may uncover differentiation even when an electrophoretic analysis suggests
that genetic differentiation does not occur (link 1991, Karl and Avise 1992).

Regardless of whether other molecular techniques reveal latitudinal
variation in P. glaucus, the electrophoretic analysis suggests that gene flow is
sufficiently common throughout the range of this butterfly so as to act as a
homogenizing influence. In a heterogeneous environment, gene flow would
counteract differential selection on different hosts thereby drastically slowing the
rate of optimization across hosts. However, even in a heterogeneous
environment, optimization would be expected to evolve eventually such that
genotypes with high fitness on all hosts would predominate, given stable
environmental conditions and equal gene flow among hosts. The observation that
optimization has not evolved in P. glaucus in over 20,000 generations since
glaciation, implies that either tradeoffs are present or conditions are not at
equilibrium, or both. Both of these factors probably define host use patterns in
P. glaucus. Though weak and indirect, evidence of tradeoffs in host use is
present. Moreover, host suitability likely changes both temporally and spatially.
Optimum genotypes are probably dynamic, reflecting a stochastic, constantly
changing host environment. Equilibirium conditions, in general, are probably very
rare in natural populations. I predict that differential selection on ecologically
important variation in P. glaucus will continue to be countered by gene flow,
thereby maintaining genetic variation for host-use traits. In the absence of more
substantial barriers to gene flow, the eastern tiger swallowtail will continue to be
comprised of a mosaic of genotypes exhibiting differing abilities to use different

host species.

APPENDIX

93
APPENDIX 1

Record of Deposition of Voucher Specimens*

The specimens listed on the following sheet(s) have been deposited in
the named museum(s) as samples of those species or other taxa which were
used in this research. Voucher recognition labels bearing the Voucher
No. have been attached or included in fluid-preserved specimens.

Voucher No.: 1993-1
Title of thesis or dissertation (or other research projects):
Differential selection and adaptation in different host environments:

genotypic and phenotypic variation in host use traits in the tiger
swallowtail butterfly, Papilio glaucus L.

 

Museum(s) where deposited and abbreviations for table on following sheets:

Entomology Museum, Michigan State University (MSU)

Other Museums:

Investigator's Name (5) (typed)
Janice L. Bossart

 

 

Date 7 June 1993

*Reference: Yoshimoto, C. M. 1978. Voucher Specimens for Entomology in
North America. Bull. Entomol. Soc. Amer. 24:141-42.

Deposit as follows:

Original: Include as Appendix 1 in ribbon copy of thesis or
dissertation.
Copies: Included as Appendix 1 in copies of thesis or dissertation.

Museum(s) files.
Research project files.

This form is available from and the Voucher No. is assigned by the Curator,
Michigan State University Entomology Museum.

94
APPENDIX 1.1

Voucher Specimen Data

Pages

_l___

of

.1...

Page

mama

“cumuso

 

mufimum>wca wuwum amwwcofiz may a“ uwmoaww

 

mama w::w.~ mums

 

 

.Esomaz honoEoucm

 

pom mcmsaumam cwumaa m>onm may vm>wmumm

 

4-mmm4 .oz amsuso>

“Lennon .4 .w
Avoahuv vamsmz m4u0umwfiumm>cH

Axummmoooc ma mummcm Hmcowufivvm omsv

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

mmca< .a .z
Hmmfi am:
am: e .ou mucmc3m4 "Io .4 mausoFm owpwmma
ucommom .4 .n
mmmfi umcmmc am4
pm: A .00 mocwczm4 "Io .4 mzuzmpm owpwgma
xmpmcamz .w
mwafl sm=m=<
:m: m .ou memFQ "<w .4 mzuzmpm owpwamm
Hemmmom .4 .n
mmmfi umgmmc nm4
am: o .oo exempu ”<w .4 muons—m owpmama
acmmmom .4 .a
mmas _cta<
2m: m a .00 muca_;mwz "4u .4 asusapm owp_ama
Juo+ woufimoawv vcm com: no vmuumHHoo coxmu umnuo no mowumam
m e y r m m e .m % mewEwomam new mumv Hmong
erOd€llapWS
s e D.& .n u u D. m "5
uhettdduyag
deiOAAPNLE c
"mo uwpssz

BIBLIOGRAPHY

BIBLIOGRAPHY

Arnold, S. 1987. Genetic correlation and the evolution of physiology. In: M. D.
Feder, A. F. Bennett, W. W. Burggren and R. B. Huey (eds.). New
directions in ecological physiology. Cambridge Univ. Press. New York.
pp. 189-215.

Ayres, M. P., and D. L. Thomas. 1990. Alternative formulations of the mixed-
model ANOVA applied to quantitative genetics. Evolution. 44:221-226.

Ayres, M. P., J. L. Bossart, and J. M. Scriber. 1991. Variation in the nutritional
physiology of tree-feeding swallowtail caterpillars. In: Y. N. Bamachikov,
W. J. Mattson, F. P. Hain, and T. L. Payne (eds.). Forest insect guilds:
patterns of interaction with host trees. USDA, Forest Service, Radnor, Pa.
pp. 85-102.

Becker, W. A. 1984. Manual of quantitative genetics. 4th edition. Academic
Enterprises. Pullman, Washington. 188 pp.

Berenbaum, M. 1981. Note on an oviposition "mistake" by Papilio glaucus
(Papilionidae). J. Lep. Soc. 35:75.

Bernays, E. A., and M. Graham. 1988. On the evolution of host specificity in
phytophagous arthropods. Ecology. 69:886-892.

Blau, W. S. 1981. Life histroy variation in the black swallowtail butterfly.
Oecologia. 48:116-122.

Bossart, J. L., and J. M. Scriber. 1993. Genetic variation in oviposition preference
in tiger swallowtail butterflies: Interspecific, interpopulation and
interindividual comparisons. In: J. M. Scriber, Y. Tsubaki and R. C.
Lederhouse (eds.). The swallowtail butterflies: Their ecology and
evolutionary biology. Scientific Publishers, Inc. Gainesville, FL. In press.

Corbet, S. A. 1985. Insect chemosensory responses: a chemical legacy hypothesis.
Ecol. Entomol. 10:143-153.

Courtney S. P., G. K. Chen, and A. Gardner. 1989. A general model for
individual host selection. Oikos. 55:55-65.

95

96

Ehrlich, P. R., and P. H. Raven. 1969. Differentiation of populations. Science.
165:1228-1232

Falconer, D. S. 1981. Introduction to quantitative genetics. 2nd ed. Longman.
New York. 340 pp.

Feeny, P. 1991. Chemical constraints on the evolution of swallowtail butterflies.
In: P. W. Price, T. M. Lewinsohn, G. W. Fernandes and W. W. Benson
(eds.). Plant-animal interactions: Evolutionary ecology in tropical and
temperate regions. Wiley & Sons, New York. 639 pp.

Fox, L. R., and P. A. Morrow. 1981. Specialization: species property or local
phenomenon? Science. 211:887-893.

Fry, J. D. 1990. Trade-offs in fitness on different hosts: evidence from a selection
experiment with a phytophagous mite. Am. Nat. 136:569-580.

Fry, J. D. 1993. The "general vigor" problem: can antagonistic pleiotropy be
detected when genetic covariances are positive? Evolution. 47:327-333.

Futuyma, D. J. 1986. Evolutionary Biology. Sinauer Assoc. Sunderland, Mass. 565 J
PP-

Futuyma, D. J. 1991. Evolution of host specificity in herbivorous insects: genetic,
ecological and phylogenetic aspects. In: P. W. Price, T. M. Lewinsohn,
G. W. Fernandes, and W. W. Benson (eds.). Plant-animal interactions:
evolutionary ecology in tropical and temperate regions. Wiley & Sons,
Inc. pp. 431-454.

Futuyma, D. J., and S. C. Peterson. 1985. Genetic variation in the use of
resources by insects. Ann. Rev. Entomol. 30:217-38.
Gilbert, L. E., and M. C. Singer. 1975. Butterfly ecology. Ann. Rev. Ecol. Syst. r?
6:365-397. it. -
I

Gillespie, J. H. 1973. Polymorphism in random environments. Theo. Pop. Biol. '.
4:193-195. ;

Gillespie, J. H., and M. Turelli. 1989. Genotype-environment interactions and the
maintenance of polygenic variation. Genetics. 121:129-138.

Gould, F. 1979. Rapid host range evolution in a population of the phytophagous
mite Tetranychus urticae Koch. Evolution. 33:791-802.

Grossmueller, D. W., and R. C. Lederhouse. 1985. The role of nectar source
distribution in habitat use and oviposition by the tiger swallowtail
butterfly. J. Lep. Soc. 41:159-165.

97

Hagen, R. H., and J. M. Scriber. 1989. Sex linked diapause, color and allozyme
loci in Papilio glaucus: linkage analysis and significance in a hybrid zone.
Heredity. 80:179-185.

Hagen, R. H., and J. M. Scriber. 1991. Systematics of the Papilio glaucus and
P. troilus species groups (Lepidoptera: Papilionidae): Inferences from
allozymes. Ann. Ent. Soc. Am. 84:380-395.

Hagen, R. H., R. C. Lederhouse, J. I... Bossart, and J. M. Scriber. 1991. Papilio
canadensis and P. glaucus (Papilionidae) are distinct species. J. Lep. Soc.
45:245-258.

Harris, H., and D. A. Hopkinson. 1978. Handbook of enzyme electrophoresis in
human genetics. American Elsevier, New York.

Haukioja, E., and S. Neuvonen. 1985. The relationship between size and
reproductive potential in male and female Epim'ta autumnata
(Lepidoptera: Geometridae). Ecological Entomology. 10:267-270.

Hedrick, P. O., M. E. Ginevan, and E. P. Ewing. 1976. Genetic polymorphism in
heterogeneous environments. Ann. Rev. Ecol. Syst. 7:1-32.

Hedrick, P. O. 1986. Genetic polymorphism in heterogeneous environments: a
decade later. Ann. Rev. Ecol. Syst. 17:535-566.

Hopkins, A. D. 1917. A discussion of C. G. Hewitt’s paper on Insect Behaviour.
J. Econ. Entomol. 10:92-93.

Jaenike, J. 1983. Induction of host preference in Drosophila melanogaster.
Oecologia 58:320-325.

Jaenike, J. 1988. Effects of early experience on host selection in insects: some
experimental and theoretical results. J. Insect Behav. 1:3-15.

Jaenike, J. 1989. Genetic population structure of Drosophila tn’punctata: patterns
of variation and covariation of traits affecting resource use. Evolution.
43:1467-1482.

Jaenike, J. 1990. Host specialization in phytophagous insects. Ann. Rev. Ecol.
Syst. 21:243-273.

Jaenike, J., and R. D. Holt. 1991. Genetic variation for habitat preference:
evidence and explanations. Am. Nat. 137:567-590.

Jaenike, J., and D. R. Papaj. 1992. Behavioral plasticity and patterns of host use
by insects. In: B. D. Roitberg and M. B. Isman (eds.). Insect chemical
ecology: an evolutionary approach. Chapman and Hall. New York. pp.
245-264.

98

Johnson, M. S., B. Clarke, and J. Murray. 1988. Discrepancies in the estimation of
gene flow in Partula. Genetics. 120:233-238.

Juniper, B., and R. Southwood. 1986. Insects and the Plant Surface. Edward
Arnold Pub. London. 360 pp.

Karl, S. A., and J. C. Avise. 1992. Balancing selection at allozyme loci in oysters:
implications from nuclear RFLP’s. Science. 256:100-102.

Keohn, R. K., and T. J. Hilbish. 1987. The adaptive importance of genetic
variation. Amer. Sci. 75:134-141.

Lande, R. 1984. The genetic correlation between characters maintained by
selection, linkage and inbreeding. Genet. Res. 44:309-320.

Lederhouse, R. C. 1983. Population structure, residency and weather related
mortality in the black swallowtail butterfly, Papilio polyxenes. Oecologia.
59:307-311.

Lederhouse, R. C., and J. M. Scriber. 1987. Increased relative frequency of dark
morph females in the tiger swallowtail Papilio glaucus (Lepidoptera:
Papilionidae) in south-central Florida. Am. Mid. Nat. 118:211-213

Lederhouse, R. C., M. P. Ayres, and J. M. Scriber. 1990. Adult nutrition affects
male virility in Papilio glaucus L. Functional Ecology. 4:743-751.

Lehman, E. L. 1975. Nonparametrics: statistical methods based on ranks. Holden-
Day, Inc., Oakland, California. 457 pp.

Levene, H. 1953. Genetic equilibrium when more than one ecological niche is
available. Am. Nat. 87:331-333.

Lindroth, R. 1988. Hydrolysis of phenolic glycosides by midgut B-glucosidases in
Papilio glaucus subspecies. Insect Biochem. 8:789-792.

Lindroth, R. L., J. M. Scriber, and M. T. S. Hsia. 1986. Differential responses of
tiger swallowtail subspecies to secondary metabolites from tuliptree and
quaking aspen. Oecologia. 70:13-19.

Long, J. C. 1986. The allelic correlation structure of Gainj- and Kalam-speaking
people. 1. The estimation and interpretation of Wright’s F-statistics.
Genetics. 112:629-647.

Luebeke, H. J., J. M. Scriber, and B. S. Yandell. 1988. Use of multivariate
discriminant analysis of male wing morphometrics to delineate the
Wisconsin hybrid zone for Papilio glaucus glaucus and P. g. canadensis.
Amer. Midl. Nat. 119:366-379.

99

Mangel M. 1989. An evolutionary interpretation of the "motivation to oviposit".\’J.
Evol. Biol. 2:157-172.

Mattson, W. J. 1980. Herbivory in relation to plant nitrogen content. Ann. Rev.
Ecol. Syst. 11:119-161.

Maynard Smith, J. 1966. Sympatric speciation. Am. Nat. 100:637-650.

McCauley, D. E., and W. F. Eanes. 1987. Hierarchical population structure
analysis of the milkweed beetle, T etraopes tetraophthalmus (Forster).
Heredity. 58:193-201.

Michaud, J. P. 1990. Conditions for the evolution of polyphagy in herbivorous
insects. Oikos 57:278-279.

Minkenberg, O. P. J. M., M. Tatar, and J. A. Rosenheim. 1992. Egg load as a
major source of variability in insect foraging and oviposition behavior.
Oikos 65:134-143.

Mitchell-Olds, T. 1992. Does environmental variation maintain genetic variation?
A question of scale. TREE. 12:397-398.

Nevo, E. 1988. Genetic differentiation in evolution. In: 181 Atlas of Science:
Animal and Plant Sciences. pp. 195-202.

Ng, D. 1988. A novel level of interactions in plant-insect systems. Nature 334:611-
612.

Nitao, J. K., M. P. Ayres, R. C. Lederhouse, and J. M. Scriber. 1991a. Larval
adaptation to lauraceous hosts: geographic divergence in the spicebush
swallowtail butterfly. Ecology. 72: 1428-1435.

Nitao, J. K., M. G. Nair, D. L. Thorogood, K. S. Johnson, and J. M. Scriber.
1991b. Bioactive neolignans from the leaves of Magnolia virginiana.
Phytochemistry. 30:2193-2195.

Odendaal, F. J ., and M. D. Rausher. 1990. Egg load influences search intensity,
host selectivity, and clutch size in Battus philenor butterflies. J. Insect
Behav. 3:183-193.

Powell, J. R. 1971. Genetic polymorphisms in varied environments. Science.
174:1035-1036.

Powell, J. R., and C. E. Taylor. 1979. Genetic variation in ecologically diverse
environments. Am. Scientist. 67:590-596.

100

Rausher, M. D. 1983. Conditioning and genetic variation as causes of individual
variation in the oviposition behavior of the tortoise beetle, Deloyala
guttata. Anim. Behav. 31:743-747.

Rausher, M. D. 1984. Tradeoffs in performance on different hosts: evidence from
within-and between-site variation in the beetle Deloyala guttata.
Evolution. 38:582-595.

Rockey, S. J., J. H. Hainze, and J. M. Scriber. 1987a. A latitudinal and obligatory
diapause response in three subspecies of the eastern tiger swallowtail,
Papilio glaucus (Lepidoptera: Papilionidae). Amer. Midl. Nat. 118:162-
168.

Rockey, S. J., J. H. Hainze, and J. M. Scriber. 1987b. Evidence of a sex-linked
diapause response in Papilio glaucus subspecies and their hybrids. Phys.
Entomol. 12:181-184.

Rose, M. 1983. Further models of selection with antagonistic pleiotropy. In: H. I.
Freedman and C. Strobeck (eds.). Population biology. Springer-Verlan.
New York. pp. 47-53.

Richardson, B. J., P. R. Baverstock, and M. Adams. 1986. Allozyme
electrophoresis. Academic, Sydney.

Rosenthal, G. A., and D. H. Janzen. 1979. Herbivores: Their Interaction with
Secondary Plant Metabolites. Academic Press. New York, London. 718

PP-

Rossiter, M. C. 1987. Genetic and phenotypic variation in diet breadth in a
generalist herbivore. Evolutionary Ecology. 1:272-282.

Rossiter, M. C. 1991a. Maternal effects generate variation in life history:
consequences of egg weight plasticity in the gypsy moth. Funct. Ecol.
5:386-393.

Rossiter, M. C. 1991b. ENvironmentally-based maternal effects: a hidden force in
insect population dynamics? Oecologia. 87:288-294.

SAS Institute Inc. 1985. SAS User’s Guide: statistics. 5th edition. 956 pp.

Scriber, J. M. 1982. Foodplants and speciation in the Papilio glaucus group. In: J.
H. Visser and A. K. Minks (eds.). Proc. 5th international symposium on
insect plant relationships. PUDOC. Wageningen, Netherlands. pp. 307-
314.

101

Scriber, J. M. 1983. Evolution of feeding specialization, physiological efficiency,
and host races in selected Papilionidae and Saturniidae. In R. F. Denno,
and M. S. McClure (eds.). Variable plants and herbivores in natural and
managed systems. Academic Press, New York. pp. 373-412.

Scriber, J. M. 1984. Larval food plant utilization by the world Papilionidae
(Lepidoptera): latitudinal gradients reappraised. Tokurana (Acta
Rhopaloc.). 2:1-50.

Scriber, J. M. 1986. Origins of the regional feeding abilities in the tiger
swallowtail butterfly: ecological monophagy and the Papilio glaucus
australis subspecies in Florida. Oecologia. 71:94-103.

Scriber, J. M. 1988. Tale of the tiger: Beringial biogeography, binomial
classification, and breakfast choices in the Papilio glaucus complex of
butterflies. In: K. C. Spencer (ed.). Chemical Mediation of Coevolution.
Academic Press. New York. pp. 241-301.

Scriber, J. M. 1993. Absence of behavioral induction in multi-choice oviposition
preference studies with a generalist butterfly species, Papilio glaucus.
Great Lakes Ent. In press.

Scriber, J. M., and R. C. Lederhouse. 1992. The thermal environment as a
resource dictating geographic patterns of feeding specialization of insect
herbivores. In: M. R. Hunter, R. Ohgushi and P. W. Price (eds.). Effects
of resource distribution on animal-plant interactions. Academic Press.
New York. pp. 429-466.

Scriber, J. M., and F. Slansky Jr. 1981. The nutritional ecology of immature
insects. Ann. Rev. Entomol. 26:183-211.

Scriber, J. M., R. L. lindroth, and J. Nitao. 1989. Differential toxicity of a
phenolic glycoside from quaking aspen to Papilio glaucus butterfly
subspecies, hybrids and backcrosses. Oecologia. 81:186-191.

Scriber, J. M., R. C. Lederhouse, and R. C. Hagen. 1991a. Foodplants and
evolution within the Papilio glaucus and P. troilus species groups
(Ledpidoptera: Papilioniade). In: P. W. Price, T. M. Lewinsohn, G. W.
Femandes, and W. W. Benson (eds.). Plant-animal interactions:
evolutionary ecology in tropical and temperate regions. Wiley and Sons,
Inc. pp. 341-373.

Scriber, J. M., B. L. Giebink, and D. Snider. 1991b. Reciprocal latitudinal clines
in oviposition behavior of Papilio glaucus and P. canadensis across the

Great Lakes hybrid zone: possible sex-linkage of oviposition preferences.
Oecologia. 87:360-368.

102

Singer, M. C. 1986. The definition and measurement of oviposition preference. In:
J. Miller and T. A. Miller (eds.). Plant-insect interactions. Springer-
Verlag. pp. 65-94.

Singer, M. C., D. Ng, and C. D. Thomas. 1988. Heritability of oviposition
preference and its relationship to offspring performance within a single
insect population. Evolution 42:977-985.

Singer, M. C., D. Ng, D. Vasco, and C. D. Thomas. 1992. Rapidly evolving

associations among oviposition preferences fail to constrain evolution of
insect diet. Am. Nat. 193:9-20.

Slatkin, M. 1973. Gene flow and selection in a cline. Genetics. 75:733-756.

Slatkin, M. 1985. Gene flow in natural populations. Ann. Rev. Ecol. Syst. 16:393-
430.

Slatkin, M. 1987. Gene flow and the geographic structure of natural populations.
Science. 236:787-792.

Sperling, F. 1991. Mitochondrial DNA phylogeny, speciation and hostplant

coevolution of Papilio butterflies. Ph.D. Dissertation. Cornell University,
Ithaca, NY. 132 pp.

Sokal, R. R. and F. J. Rohlf. Biometry. W. H. Freeman & Co. San Francisco. 859
PP-

Stearns, S., G. de Long and B. Newman. 1991. The effects of phenotypic plasticity
on genetic correlations. TREE. 6:122-126.

Strong, D. R., J. H. Lawton, and R. Southwood. 1984. Insects on plants:

Community patterns and mechanisms. Harvard Univ. Press. Cambridge,
Mass.

Tabashnik, B. E., H. Wheelock, J. D. Rainbolt, and W. B. Watt. 1981. Individual

variation in oviposition preference in the butterfly, Colias ewytheme.
Oecologia 50:225-230.

Thompson, J. N. 1988. Variation in preference and specificity in monophagous
and oligophagous swallowtail butterflies. Evolution. 42:118-128.

Via, S. 1984. The quantitative genetics of polyphagy in an insect herbivore. II.

Genetic correlations in larval performance within and among host plants.
Evolution. 38:896-905. '

Via, S., and R. Lande. 1985. Genotype-environment interaction and the evolution
of phenotypic plasticity. Evolution. 39:505-521.

103

Via, S. 1986. Genetic covariance between oviposition preference and larval
performance in an insect herbivore. Evolution 40:778-785.

Via, S. 1990. Ecological genetics and host adaptation in herbivorous insects: the
experimental study of evolution in natural and agricultural systems. Ann.
Rev. Entomol. 35:421-446.

Wasserman, S. S. 1986. Genetic variation in adaptation to foodplants among
populations of the southern cowpea weevil, Callosobruchus maculatus:
evolution of oviposition preference. Entomol. Exp. Appl. 42:201-212.

Wcislo, W. T. 1989. Behavioral environments and evolutionary change. Ann. Rev.
Ecol. Syst. 20: 137-169.

Weir, B., and C. C. Cockerham. 1984. Estimating F-statistics for the analysis of
population structure. Evolution. 18:1358- 1370.

Whitlock, M. C. 1992. Temporal fluctuations in demographic parameters and the
genetic variance among populations. Evolution. 46:608-615.

Wildund, C. 1973. Host plant suitability and the mechanism of host selection in
larvae of Papilio machaon. Entomol. Exp. Appl. 16:232-242.

Wiklund, C. 1975. The evolutionary relationship between adult oviposition
preferences and larval host plant range in Papilio machaon L. Oecologia.
18:185-197.

Wright, S. 1951. The genetical structure of populations. Ann. Eugen. 15:323-354.

Zink, R. M. 1991. The geography of mitochondrial DNA variation in two
sympatric sparrows. Evolution. 45:329-339.

"1111111111111“